diff --git a/daemons/fenced/fenced_commands.c b/daemons/fenced/fenced_commands.c index bfb0d71e5f..975f8633a4 100644 --- a/daemons/fenced/fenced_commands.c +++ b/daemons/fenced/fenced_commands.c @@ -1,3256 +1,3276 @@ /* * Copyright 2009-2021 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include GHashTable *device_list = NULL; GHashTable *topology = NULL; GList *cmd_list = NULL; struct device_search_s { /* target of fence action */ char *host; /* requested fence action */ char *action; /* timeout to use if a device is queried dynamically for possible targets */ int per_device_timeout; /* number of registered fencing devices at time of request */ int replies_needed; /* number of device replies received so far */ int replies_received; /* whether the target is eligible to perform requested action (or off) */ bool allow_suicide; /* private data to pass to search callback function */ void *user_data; /* function to call when all replies have been received */ void (*callback) (GList * devices, void *user_data); /* devices capable of performing requested action (or off if remapping) */ GList *capable; }; static gboolean stonith_device_dispatch(gpointer user_data); static void st_child_done(int pid, const pcmk__action_result_t *result, void *user_data); static void stonith_send_reply(xmlNode * reply, int call_options, const char *remote_peer, const char *client_id); static void search_devices_record_result(struct device_search_s *search, const char *device, gboolean can_fence); static int get_agent_metadata(const char *agent, xmlNode **metadata); static void read_action_metadata(stonith_device_t *device); typedef struct async_command_s { int id; int pid; int fd_stdout; int options; int default_timeout; /* seconds */ int timeout; /* seconds */ int start_delay; /* seconds */ int delay_id; char *op; char *origin; char *client; char *client_name; char *remote_op_id; char *victim; uint32_t victim_nodeid; char *action; char *device; GList *device_list; GList *device_next; void *internal_user_data; void (*done_cb) (int pid, const pcmk__action_result_t *result, void *user_data); guint timer_sigterm; guint timer_sigkill; /*! If the operation timed out, this is the last signal * we sent to the process to get it to terminate */ int last_timeout_signo; stonith_device_t *active_on; stonith_device_t *activating_on; } async_command_t; static xmlNode *construct_async_reply(async_command_t *cmd, const pcmk__action_result_t *result); static gboolean is_action_required(const char *action, stonith_device_t *device) { return device && device->automatic_unfencing && pcmk__str_eq(action, "on", pcmk__str_casei); } static int get_action_delay_max(stonith_device_t * device, const char * action) { const char *value = NULL; int delay_max = 0; if (!pcmk__is_fencing_action(action)) { return 0; } value = g_hash_table_lookup(device->params, PCMK_STONITH_DELAY_MAX); if (value) { delay_max = crm_parse_interval_spec(value) / 1000; } return delay_max; } static int get_action_delay_base(stonith_device_t *device, const char *action, const char *victim) { char *hash_value = NULL; int delay_base = 0; if (!pcmk__is_fencing_action(action)) { return 0; } hash_value = g_hash_table_lookup(device->params, PCMK_STONITH_DELAY_BASE); if (hash_value) { char *value = strdup(hash_value); char *valptr = value; CRM_ASSERT(value != NULL); if (victim) { for (char *val = strtok(value, "; \t"); val != NULL; val = strtok(NULL, "; \t")) { char *mapval = strchr(val, ':'); if (mapval == NULL || mapval[1] == 0) { crm_err("pcmk_delay_base: empty value in mapping", val); continue; } if (mapval != val && strncasecmp(victim, val, (size_t)(mapval - val)) == 0) { value = mapval + 1; crm_debug("pcmk_delay_base mapped to %s for %s", value, victim); break; } } } if (strchr(value, ':') == 0) { delay_base = crm_parse_interval_spec(value) / 1000; } free(valptr); } return delay_base; } /*! * \internal * \brief Override STONITH timeout with pcmk_*_timeout if available * * \param[in] device STONITH device to use * \param[in] action STONITH action name * \param[in] default_timeout Timeout to use if device does not have * a pcmk_*_timeout parameter for action * * \return Value of pcmk_(action)_timeout if available, otherwise default_timeout * \note For consistency, it would be nice if reboot/off/on timeouts could be * set the same way as start/stop/monitor timeouts, i.e. with an * entry in the fencing resource configuration. However that * is insufficient because fencing devices may be registered directly via * the fencer's register_device() API instead of going through the CIB * (e.g. stonith_admin uses it for its -R option, and the executor uses it * to ensure a device is registered when a command is issued). As device * properties, pcmk_*_timeout parameters can be grabbed by the fencer when * the device is registered, whether by CIB change or API call. */ static int get_action_timeout(stonith_device_t * device, const char *action, int default_timeout) { if (action && device && device->params) { char buffer[64] = { 0, }; const char *value = NULL; /* If "reboot" was requested but the device does not support it, * we will remap to "off", so check timeout for "off" instead */ if (pcmk__str_eq(action, "reboot", pcmk__str_casei) && !pcmk_is_set(device->flags, st_device_supports_reboot)) { crm_trace("%s doesn't support reboot, using timeout for off instead", device->id); action = "off"; } /* If the device config specified an action-specific timeout, use it */ snprintf(buffer, sizeof(buffer), "pcmk_%s_timeout", action); value = g_hash_table_lookup(device->params, buffer); if (value) { return atoi(value); } } return default_timeout; } static void free_async_command(async_command_t * cmd) { if (!cmd) { return; } if (cmd->delay_id) { g_source_remove(cmd->delay_id); } cmd_list = g_list_remove(cmd_list, cmd); g_list_free_full(cmd->device_list, free); free(cmd->device); free(cmd->action); free(cmd->victim); free(cmd->remote_op_id); free(cmd->client); free(cmd->client_name); free(cmd->origin); free(cmd->op); free(cmd); } static async_command_t * create_async_command(xmlNode * msg) { async_command_t *cmd = NULL; xmlNode *op = get_xpath_object("//@" F_STONITH_ACTION, msg, LOG_ERR); const char *action = crm_element_value(op, F_STONITH_ACTION); CRM_CHECK(action != NULL, crm_log_xml_warn(msg, "NoAction"); return NULL); crm_log_xml_trace(msg, "Command"); cmd = calloc(1, sizeof(async_command_t)); crm_element_value_int(msg, F_STONITH_CALLID, &(cmd->id)); crm_element_value_int(msg, F_STONITH_CALLOPTS, &(cmd->options)); crm_element_value_int(msg, F_STONITH_TIMEOUT, &(cmd->default_timeout)); cmd->timeout = cmd->default_timeout; // Value -1 means disable any static/random fencing delays crm_element_value_int(msg, F_STONITH_DELAY, &(cmd->start_delay)); cmd->origin = crm_element_value_copy(msg, F_ORIG); cmd->remote_op_id = crm_element_value_copy(msg, F_STONITH_REMOTE_OP_ID); cmd->client = crm_element_value_copy(msg, F_STONITH_CLIENTID); cmd->client_name = crm_element_value_copy(msg, F_STONITH_CLIENTNAME); cmd->op = crm_element_value_copy(msg, F_STONITH_OPERATION); cmd->action = strdup(action); cmd->victim = crm_element_value_copy(op, F_STONITH_TARGET); cmd->device = crm_element_value_copy(op, F_STONITH_DEVICE); CRM_CHECK(cmd->op != NULL, crm_log_xml_warn(msg, "NoOp"); free_async_command(cmd); return NULL); CRM_CHECK(cmd->client != NULL, crm_log_xml_warn(msg, "NoClient")); cmd->done_cb = st_child_done; cmd_list = g_list_append(cmd_list, cmd); return cmd; } static int get_action_limit(stonith_device_t * device) { const char *value = NULL; int action_limit = 1; value = g_hash_table_lookup(device->params, PCMK_STONITH_ACTION_LIMIT); if ((value == NULL) || (pcmk__scan_min_int(value, &action_limit, INT_MIN) != pcmk_rc_ok) || (action_limit == 0)) { action_limit = 1; } return action_limit; } static int get_active_cmds(stonith_device_t * device) { int counter = 0; GList *gIter = NULL; GList *gIterNext = NULL; CRM_CHECK(device != NULL, return 0); for (gIter = cmd_list; gIter != NULL; gIter = gIterNext) { async_command_t *cmd = gIter->data; gIterNext = gIter->next; if (cmd->active_on == device) { counter++; } } return counter; } static void fork_cb(int pid, void *user_data) { async_command_t *cmd = (async_command_t *) user_data; stonith_device_t * device = /* in case of a retry we've done the move from activating_on to active_on already */ cmd->activating_on?cmd->activating_on:cmd->active_on; CRM_ASSERT(device); crm_debug("Operation '%s' [%d]%s%s using %s now running with %ds timeout", cmd->action, pid, ((cmd->victim == NULL)? "" : " targeting "), ((cmd->victim == NULL)? "" : cmd->victim), device->id, cmd->timeout); cmd->active_on = device; cmd->activating_on = NULL; } static int get_agent_metadata_cb(gpointer data) { stonith_device_t *device = data; guint period_ms; switch (get_agent_metadata(device->agent, &device->agent_metadata)) { case pcmk_rc_ok: if (device->agent_metadata) { read_action_metadata(device); stonith__device_parameter_flags(&(device->flags), device->id, device->agent_metadata); } return G_SOURCE_REMOVE; case EAGAIN: period_ms = pcmk__mainloop_timer_get_period(device->timer); if (period_ms < 160 * 1000) { mainloop_timer_set_period(device->timer, 2 * period_ms); } return G_SOURCE_CONTINUE; default: return G_SOURCE_REMOVE; } } /*! * \internal * \brief Call a command's action callback for an internal (not library) result * * \param[in] cmd Command to report result for * \param[in] execution_status Execution status to use for result * \param[in] exit_status Exit status to use for result * \param[in] exit_reason Exit reason to use for result */ static void report_internal_result(async_command_t *cmd, int exit_status, int execution_status, const char *exit_reason) { pcmk__action_result_t result = PCMK__UNKNOWN_RESULT; pcmk__set_result(&result, exit_status, execution_status, exit_reason); cmd->done_cb(0, &result, cmd); pcmk__reset_result(&result); } static gboolean stonith_device_execute(stonith_device_t * device) { int exec_rc = 0; const char *action_str = NULL; const char *host_arg = NULL; async_command_t *cmd = NULL; stonith_action_t *action = NULL; int active_cmds = 0; int action_limit = 0; GList *gIter = NULL; GList *gIterNext = NULL; CRM_CHECK(device != NULL, return FALSE); active_cmds = get_active_cmds(device); action_limit = get_action_limit(device); if (action_limit > -1 && active_cmds >= action_limit) { crm_trace("%s is over its action limit of %d (%u active action%s)", device->id, action_limit, active_cmds, pcmk__plural_s(active_cmds)); return TRUE; } for (gIter = device->pending_ops; gIter != NULL; gIter = gIterNext) { async_command_t *pending_op = gIter->data; gIterNext = gIter->next; if (pending_op && pending_op->delay_id) { crm_trace("Operation '%s'%s%s using %s was asked to run too early, " "waiting for start delay of %ds", pending_op->action, ((pending_op->victim == NULL)? "" : " targeting "), ((pending_op->victim == NULL)? "" : pending_op->victim), device->id, pending_op->start_delay); continue; } device->pending_ops = g_list_remove_link(device->pending_ops, gIter); g_list_free_1(gIter); cmd = pending_op; break; } if (cmd == NULL) { crm_trace("No actions using %s are needed", device->id); return TRUE; } if (pcmk__str_any_of(device->agent, STONITH_WATCHDOG_AGENT, STONITH_WATCHDOG_AGENT_INTERNAL, NULL)) { if (pcmk__is_fencing_action(cmd->action)) { if (node_does_watchdog_fencing(stonith_our_uname)) { pcmk__panic(__func__); goto done; } } else { crm_info("Faking success for %s watchdog operation", cmd->action); report_internal_result(cmd, CRM_EX_OK, PCMK_EXEC_DONE, NULL); goto done; } } #if SUPPORT_CIBSECRETS exec_rc = pcmk__substitute_secrets(device->id, device->params); if (exec_rc != pcmk_rc_ok) { if (pcmk__str_eq(cmd->action, "stop", pcmk__str_casei)) { crm_info("Proceeding with stop operation for %s " "despite being unable to load CIB secrets (%s)", device->id, pcmk_rc_str(exec_rc)); } else { crm_err("Considering %s unconfigured " "because unable to load CIB secrets: %s", device->id, pcmk_rc_str(exec_rc)); report_internal_result(cmd, CRM_EX_ERROR, PCMK_EXEC_NO_SECRETS, "Failed to get CIB secrets"); goto done; } } #endif action_str = cmd->action; if (pcmk__str_eq(cmd->action, "reboot", pcmk__str_casei) && !pcmk_is_set(device->flags, st_device_supports_reboot)) { crm_notice("Remapping 'reboot' action%s%s using %s to 'off' " "because agent '%s' does not support reboot", ((cmd->victim == NULL)? "" : " targeting "), ((cmd->victim == NULL)? "" : cmd->victim), device->id, device->agent); action_str = "off"; } if (pcmk_is_set(device->flags, st_device_supports_parameter_port)) { host_arg = "port"; } else if (pcmk_is_set(device->flags, st_device_supports_parameter_plug)) { host_arg = "plug"; } action = stonith_action_create(device->agent, action_str, cmd->victim, cmd->victim_nodeid, cmd->timeout, device->params, device->aliases, host_arg); /* for async exec, exec_rc is negative for early error exit otherwise handling of success/errors is done via callbacks */ cmd->activating_on = device; exec_rc = stonith_action_execute_async(action, (void *)cmd, cmd->done_cb, fork_cb); if (exec_rc < 0) { cmd->activating_on = NULL; cmd->done_cb(0, stonith__action_result(action), cmd); stonith__destroy_action(action); } done: /* Device might get triggered to work by multiple fencing commands * simultaneously. Trigger the device again to make sure any * remaining concurrent commands get executed. */ if (device->pending_ops) { mainloop_set_trigger(device->work); } return TRUE; } static gboolean stonith_device_dispatch(gpointer user_data) { return stonith_device_execute(user_data); } static gboolean start_delay_helper(gpointer data) { async_command_t *cmd = data; stonith_device_t *device = NULL; cmd->delay_id = 0; device = cmd->device ? g_hash_table_lookup(device_list, cmd->device) : NULL; if (device) { mainloop_set_trigger(device->work); } return FALSE; } static void schedule_stonith_command(async_command_t * cmd, stonith_device_t * device) { int delay_max = 0; int delay_base = 0; int requested_delay = cmd->start_delay; CRM_CHECK(cmd != NULL, return); CRM_CHECK(device != NULL, return); if (cmd->device) { free(cmd->device); } if (device->include_nodeid && cmd->victim) { crm_node_t *node = crm_get_peer(0, cmd->victim); cmd->victim_nodeid = node->id; } cmd->device = strdup(device->id); cmd->timeout = get_action_timeout(device, cmd->action, cmd->default_timeout); if (cmd->remote_op_id) { crm_debug("Scheduling '%s' action%s%s using %s for remote peer %s " "with op id %.8s and timeout %ds", cmd->action, cmd->victim ? " targeting " : "", cmd->victim ? cmd->victim : "", device->id, cmd->origin, cmd->remote_op_id, cmd->timeout); } else { crm_debug("Scheduling '%s' action%s%s using %s for %s with timeout %ds", cmd->action, cmd->victim ? " targeting " : "", cmd->victim ? cmd->victim : "", device->id, cmd->client, cmd->timeout); } device->pending_ops = g_list_append(device->pending_ops, cmd); mainloop_set_trigger(device->work); // Value -1 means disable any static/random fencing delays if (requested_delay < 0) { return; } delay_max = get_action_delay_max(device, cmd->action); delay_base = get_action_delay_base(device, cmd->action, cmd->victim); if (delay_max == 0) { delay_max = delay_base; } if (delay_max < delay_base) { crm_warn(PCMK_STONITH_DELAY_BASE " (%ds) is larger than " PCMK_STONITH_DELAY_MAX " (%ds) for %s using %s " "(limiting to maximum delay)", delay_base, delay_max, cmd->action, device->id); delay_base = delay_max; } if (delay_max > 0) { // coverity[dont_call] We're not using rand() for security cmd->start_delay += ((delay_max != delay_base)?(rand() % (delay_max - delay_base)):0) + delay_base; } if (cmd->start_delay > 0) { crm_notice("Delaying '%s' action%s%s using %s for %ds " CRM_XS " timeout=%ds requested_delay=%ds base=%ds max=%ds", cmd->action, cmd->victim ? " targeting " : "", cmd->victim ? cmd->victim : "", device->id, cmd->start_delay, cmd->timeout, requested_delay, delay_base, delay_max); cmd->delay_id = g_timeout_add_seconds(cmd->start_delay, start_delay_helper, cmd); } } static void free_device(gpointer data) { GList *gIter = NULL; stonith_device_t *device = data; g_hash_table_destroy(device->params); g_hash_table_destroy(device->aliases); for (gIter = device->pending_ops; gIter != NULL; gIter = gIter->next) { async_command_t *cmd = gIter->data; crm_warn("Removal of device '%s' purged operation '%s'", device->id, cmd->action); report_internal_result(cmd, CRM_EX_ERROR, PCMK_EXEC_NO_FENCE_DEVICE, "Device was removed before action could be executed"); } g_list_free(device->pending_ops); g_list_free_full(device->targets, free); if (device->timer) { mainloop_timer_stop(device->timer); mainloop_timer_del(device->timer); } mainloop_destroy_trigger(device->work); free_xml(device->agent_metadata); free(device->namespace); free(device->on_target_actions); free(device->agent); free(device->id); free(device); } void free_device_list(void) { if (device_list != NULL) { g_hash_table_destroy(device_list); device_list = NULL; } } void init_device_list(void) { if (device_list == NULL) { device_list = pcmk__strkey_table(NULL, free_device); } } static GHashTable * build_port_aliases(const char *hostmap, GList ** targets) { char *name = NULL; int last = 0, lpc = 0, max = 0, added = 0; GHashTable *aliases = pcmk__strikey_table(free, free); if (hostmap == NULL) { return aliases; } max = strlen(hostmap); for (; lpc <= max; lpc++) { switch (hostmap[lpc]) { /* Skip escaped chars */ case '\\': lpc++; break; /* Assignment chars */ case '=': case ':': if (lpc > last) { free(name); name = calloc(1, 1 + lpc - last); memcpy(name, hostmap + last, lpc - last); } last = lpc + 1; break; /* Delimeter chars */ /* case ',': Potentially used to specify multiple ports */ case 0: case ';': case ' ': case '\t': if (name) { char *value = NULL; int k = 0; value = calloc(1, 1 + lpc - last); memcpy(value, hostmap + last, lpc - last); for (int i = 0; value[i] != '\0'; i++) { if (value[i] != '\\') { value[k++] = value[i]; } } value[k] = '\0'; crm_debug("Adding alias '%s'='%s'", name, value); g_hash_table_replace(aliases, name, value); if (targets) { *targets = g_list_append(*targets, strdup(value)); } value = NULL; name = NULL; added++; } else if (lpc > last) { crm_debug("Parse error at offset %d near '%s'", lpc - last, hostmap + last); } last = lpc + 1; break; } if (hostmap[lpc] == 0) { break; } } if (added == 0) { crm_info("No host mappings detected in '%s'", hostmap); } free(name); return aliases; } GHashTable *metadata_cache = NULL; void free_metadata_cache(void) { if (metadata_cache != NULL) { g_hash_table_destroy(metadata_cache); metadata_cache = NULL; } } static void init_metadata_cache(void) { if (metadata_cache == NULL) { metadata_cache = pcmk__strkey_table(free, free); } } int get_agent_metadata(const char *agent, xmlNode ** metadata) { char *buffer = NULL; if (metadata == NULL) { return EINVAL; } *metadata = NULL; if (pcmk__str_eq(agent, STONITH_WATCHDOG_AGENT_INTERNAL, pcmk__str_none)) { return pcmk_rc_ok; } init_metadata_cache(); buffer = g_hash_table_lookup(metadata_cache, agent); if (buffer == NULL) { stonith_t *st = stonith_api_new(); int rc; if (st == NULL) { crm_warn("Could not get agent meta-data: " "API memory allocation failed"); return EAGAIN; } rc = st->cmds->metadata(st, st_opt_sync_call, agent, NULL, &buffer, 10); stonith_api_delete(st); if (rc || !buffer) { crm_err("Could not retrieve metadata for fencing agent %s", agent); return EAGAIN; } g_hash_table_replace(metadata_cache, strdup(agent), buffer); } *metadata = string2xml(buffer); return pcmk_rc_ok; } static gboolean is_nodeid_required(xmlNode * xml) { xmlXPathObjectPtr xpath = NULL; if (stand_alone) { return FALSE; } if (!xml) { return FALSE; } xpath = xpath_search(xml, "//parameter[@name='nodeid']"); if (numXpathResults(xpath) <= 0) { freeXpathObject(xpath); return FALSE; } freeXpathObject(xpath); return TRUE; } #define MAX_ACTION_LEN 256 static char * add_action(char *actions, const char *action) { int offset = 0; if (actions == NULL) { actions = calloc(1, MAX_ACTION_LEN); } else { offset = strlen(actions); } if (offset > 0) { offset += snprintf(actions+offset, MAX_ACTION_LEN - offset, " "); } offset += snprintf(actions+offset, MAX_ACTION_LEN - offset, "%s", action); return actions; } static void read_action_metadata(stonith_device_t *device) { xmlXPathObjectPtr xpath = NULL; int max = 0; int lpc = 0; if (device->agent_metadata == NULL) { return; } xpath = xpath_search(device->agent_metadata, "//action"); max = numXpathResults(xpath); if (max <= 0) { freeXpathObject(xpath); return; } for (lpc = 0; lpc < max; lpc++) { const char *action = NULL; xmlNode *match = getXpathResult(xpath, lpc); CRM_LOG_ASSERT(match != NULL); if(match == NULL) { continue; }; action = crm_element_value(match, "name"); if(pcmk__str_eq(action, "list", pcmk__str_casei)) { stonith__set_device_flags(device->flags, device->id, st_device_supports_list); } else if(pcmk__str_eq(action, "status", pcmk__str_casei)) { stonith__set_device_flags(device->flags, device->id, st_device_supports_status); } else if(pcmk__str_eq(action, "reboot", pcmk__str_casei)) { stonith__set_device_flags(device->flags, device->id, st_device_supports_reboot); } else if (pcmk__str_eq(action, "on", pcmk__str_casei)) { /* "automatic" means the cluster will unfence node when it joins */ /* "required" is a deprecated synonym for "automatic" */ if (pcmk__xe_attr_is_true(match, "automatic") || pcmk__xe_attr_is_true(match, "required")) { device->automatic_unfencing = TRUE; } } if (action && pcmk__xe_attr_is_true(match, "on_target")) { device->on_target_actions = add_action(device->on_target_actions, action); } } freeXpathObject(xpath); } /*! * \internal * \brief Set a pcmk_*_action parameter if not already set * * \param[in,out] params Device parameters * \param[in] action Name of action * \param[in] value Value to use if action is not already set */ static void map_action(GHashTable *params, const char *action, const char *value) { char *key = crm_strdup_printf("pcmk_%s_action", action); if (g_hash_table_lookup(params, key)) { crm_warn("Ignoring %s='%s', see %s instead", STONITH_ATTR_ACTION_OP, value, key); free(key); } else { crm_warn("Mapping %s='%s' to %s='%s'", STONITH_ATTR_ACTION_OP, value, key, value); g_hash_table_insert(params, key, strdup(value)); } } /*! * \internal * \brief Create device parameter table from XML * * \param[in] name Device name (used for logging only) * \param[in,out] params Device parameters */ static GHashTable * xml2device_params(const char *name, xmlNode *dev) { GHashTable *params = xml2list(dev); const char *value; /* Action should never be specified in the device configuration, * but we support it for users who are familiar with other software * that worked that way. */ value = g_hash_table_lookup(params, STONITH_ATTR_ACTION_OP); if (value != NULL) { crm_warn("%s has '%s' parameter, which should never be specified in configuration", name, STONITH_ATTR_ACTION_OP); if (*value == '\0') { crm_warn("Ignoring empty '%s' parameter", STONITH_ATTR_ACTION_OP); } else if (strcmp(value, "reboot") == 0) { crm_warn("Ignoring %s='reboot' (see stonith-action cluster property instead)", STONITH_ATTR_ACTION_OP); } else if (strcmp(value, "off") == 0) { map_action(params, "reboot", value); } else { map_action(params, "off", value); map_action(params, "reboot", value); } g_hash_table_remove(params, STONITH_ATTR_ACTION_OP); } return params; } static const char * target_list_type(stonith_device_t * dev) { const char *check_type = NULL; check_type = g_hash_table_lookup(dev->params, PCMK_STONITH_HOST_CHECK); if (check_type == NULL) { if (g_hash_table_lookup(dev->params, PCMK_STONITH_HOST_LIST)) { check_type = "static-list"; } else if (g_hash_table_lookup(dev->params, PCMK_STONITH_HOST_MAP)) { check_type = "static-list"; } else if (pcmk_is_set(dev->flags, st_device_supports_list)) { check_type = "dynamic-list"; } else if (pcmk_is_set(dev->flags, st_device_supports_status)) { check_type = "status"; } else { check_type = "none"; } } return check_type; } static stonith_device_t * build_device_from_xml(xmlNode * msg) { const char *value; xmlNode *dev = get_xpath_object("//" F_STONITH_DEVICE, msg, LOG_ERR); stonith_device_t *device = NULL; char *agent = crm_element_value_copy(dev, "agent"); CRM_CHECK(agent != NULL, return device); device = calloc(1, sizeof(stonith_device_t)); CRM_CHECK(device != NULL, {free(agent); return device;}); device->id = crm_element_value_copy(dev, XML_ATTR_ID); device->agent = agent; device->namespace = crm_element_value_copy(dev, "namespace"); device->params = xml2device_params(device->id, dev); value = g_hash_table_lookup(device->params, PCMK_STONITH_HOST_LIST); if (value) { device->targets = stonith__parse_targets(value); } value = g_hash_table_lookup(device->params, PCMK_STONITH_HOST_MAP); device->aliases = build_port_aliases(value, &(device->targets)); value = target_list_type(device); if (!pcmk__str_eq(value, "static-list", pcmk__str_casei) && device->targets) { /* Other than "static-list", dev-> targets is unnecessary. */ g_list_free_full(device->targets, free); device->targets = NULL; } switch (get_agent_metadata(device->agent, &device->agent_metadata)) { case pcmk_rc_ok: if (device->agent_metadata) { read_action_metadata(device); stonith__device_parameter_flags(&(device->flags), device->id, device->agent_metadata); } break; case EAGAIN: if (device->timer == NULL) { device->timer = mainloop_timer_add("get_agent_metadata", 10 * 1000, TRUE, get_agent_metadata_cb, device); } if (!mainloop_timer_running(device->timer)) { mainloop_timer_start(device->timer); } break; default: break; } value = g_hash_table_lookup(device->params, "nodeid"); if (!value) { device->include_nodeid = is_nodeid_required(device->agent_metadata); } value = crm_element_value(dev, "rsc_provides"); if (pcmk__str_eq(value, "unfencing", pcmk__str_casei)) { device->automatic_unfencing = TRUE; } if (is_action_required("on", device)) { crm_info("Fencing device '%s' requires unfencing", device->id); } if (device->on_target_actions) { crm_info("Fencing device '%s' requires actions (%s) to be executed " "on target", device->id, device->on_target_actions); } device->work = mainloop_add_trigger(G_PRIORITY_HIGH, stonith_device_dispatch, device); /* TODO: Hook up priority */ return device; } static void schedule_internal_command(const char *origin, stonith_device_t * device, const char *action, const char *victim, int timeout, void *internal_user_data, void (*done_cb) (int pid, const pcmk__action_result_t *result, void *user_data)) { async_command_t *cmd = NULL; cmd = calloc(1, sizeof(async_command_t)); cmd->id = -1; cmd->default_timeout = timeout ? timeout : 60; cmd->timeout = cmd->default_timeout; cmd->action = strdup(action); cmd->victim = victim ? strdup(victim) : NULL; cmd->device = strdup(device->id); cmd->origin = strdup(origin); cmd->client = strdup(crm_system_name); cmd->client_name = strdup(crm_system_name); cmd->internal_user_data = internal_user_data; cmd->done_cb = done_cb; /* cmd, not internal_user_data, is passed to 'done_cb' as the userdata */ schedule_stonith_command(cmd, device); } // Fence agent status commands use custom exit status codes enum fence_status_code { fence_status_invalid = -1, fence_status_active = 0, fence_status_unknown = 1, fence_status_inactive = 2, }; static void status_search_cb(int pid, const pcmk__action_result_t *result, void *user_data) { async_command_t *cmd = user_data; struct device_search_s *search = cmd->internal_user_data; stonith_device_t *dev = cmd->device ? g_hash_table_lookup(device_list, cmd->device) : NULL; gboolean can = FALSE; free_async_command(cmd); if (!dev) { search_devices_record_result(search, NULL, FALSE); return; } mainloop_set_trigger(dev->work); if (result->execution_status != PCMK_EXEC_DONE) { crm_warn("Assuming %s cannot fence %s " "because status could not be executed: %s%s%s%s", dev->id, search->host, pcmk_exec_status_str(result->execution_status), ((result->exit_reason == NULL)? "" : " ("), ((result->exit_reason == NULL)? "" : result->exit_reason), ((result->exit_reason == NULL)? "" : ")")); search_devices_record_result(search, dev->id, FALSE); return; } switch (result->exit_status) { case fence_status_unknown: crm_trace("%s reported it cannot fence %s", dev->id, search->host); break; case fence_status_active: case fence_status_inactive: crm_trace("%s reported it can fence %s", dev->id, search->host); can = TRUE; break; default: crm_warn("Assuming %s cannot fence %s " "(status returned unknown code %d)", dev->id, search->host, result->exit_status); break; } search_devices_record_result(search, dev->id, can); } static void dynamic_list_search_cb(int pid, const pcmk__action_result_t *result, void *user_data) { async_command_t *cmd = user_data; struct device_search_s *search = cmd->internal_user_data; stonith_device_t *dev = cmd->device ? g_hash_table_lookup(device_list, cmd->device) : NULL; gboolean can_fence = FALSE; free_async_command(cmd); /* Host/alias must be in the list output to be eligible to be fenced * * Will cause problems if down'd nodes aren't listed or (for virtual nodes) * if the guest is still listed despite being moved to another machine */ if (!dev) { search_devices_record_result(search, NULL, FALSE); return; } mainloop_set_trigger(dev->work); if (pcmk__result_ok(result)) { crm_info("Refreshing target list for %s", dev->id); g_list_free_full(dev->targets, free); dev->targets = stonith__parse_targets(result->action_stdout); dev->targets_age = time(NULL); } else if (dev->targets != NULL) { if (result->execution_status == PCMK_EXEC_DONE) { crm_info("Reusing most recent target list for %s " "because list returned error code %d", dev->id, result->exit_status); } else { crm_info("Reusing most recent target list for %s " "because list could not be executed: %s%s%s%s", dev->id, pcmk_exec_status_str(result->execution_status), ((result->exit_reason == NULL)? "" : " ("), ((result->exit_reason == NULL)? "" : result->exit_reason), ((result->exit_reason == NULL)? "" : ")")); } } else { // We have never successfully executed list if (result->execution_status == PCMK_EXEC_DONE) { crm_warn("Assuming %s cannot fence %s " "because list returned error code %d", dev->id, search->host, result->exit_status); } else { crm_warn("Assuming %s cannot fence %s " "because list could not be executed: %s%s%s%s", dev->id, search->host, pcmk_exec_status_str(result->execution_status), ((result->exit_reason == NULL)? "" : " ("), ((result->exit_reason == NULL)? "" : result->exit_reason), ((result->exit_reason == NULL)? "" : ")")); } /* Fall back to pcmk_host_check="status" if the user didn't explicitly * specify "dynamic-list". */ if (g_hash_table_lookup(dev->params, PCMK_STONITH_HOST_CHECK) == NULL) { crm_notice("Switching to pcmk_host_check='status' for %s", dev->id); g_hash_table_replace(dev->params, strdup(PCMK_STONITH_HOST_CHECK), strdup("status")); } } if (dev->targets) { const char *alias = g_hash_table_lookup(dev->aliases, search->host); if (!alias) { alias = search->host; } if (pcmk__str_in_list(alias, dev->targets, pcmk__str_casei)) { can_fence = TRUE; } } search_devices_record_result(search, dev->id, can_fence); } /*! * \internal * \brief Returns true if any key in first is not in second or second has a different value for key */ static int device_params_diff(GHashTable *first, GHashTable *second) { char *key = NULL; char *value = NULL; GHashTableIter gIter; g_hash_table_iter_init(&gIter, first); while (g_hash_table_iter_next(&gIter, (void **)&key, (void **)&value)) { if(strstr(key, "CRM_meta") == key) { continue; } else if(strcmp(key, "crm_feature_set") == 0) { continue; } else { char *other_value = g_hash_table_lookup(second, key); if (!other_value || !pcmk__str_eq(other_value, value, pcmk__str_casei)) { crm_trace("Different value for %s: %s != %s", key, other_value, value); return 1; } } } return 0; } /*! * \internal * \brief Checks to see if an identical device already exists in the device_list */ static stonith_device_t * device_has_duplicate(stonith_device_t * device) { stonith_device_t *dup = g_hash_table_lookup(device_list, device->id); if (!dup) { crm_trace("No match for %s", device->id); return NULL; } else if (!pcmk__str_eq(dup->agent, device->agent, pcmk__str_casei)) { crm_trace("Different agent: %s != %s", dup->agent, device->agent); return NULL; } /* Use calculate_operation_digest() here? */ if (device_params_diff(device->params, dup->params) || device_params_diff(dup->params, device->params)) { return NULL; } crm_trace("Match"); return dup; } int stonith_device_register(xmlNode * msg, const char **desc, gboolean from_cib) { stonith_device_t *dup = NULL; stonith_device_t *device = build_device_from_xml(msg); guint ndevices = 0; int rv = pcmk_ok; CRM_CHECK(device != NULL, return -ENOMEM); /* do we have a watchdog-device? */ if (pcmk__str_eq(device->id, STONITH_WATCHDOG_ID, pcmk__str_none) || pcmk__str_any_of(device->agent, STONITH_WATCHDOG_AGENT, STONITH_WATCHDOG_AGENT_INTERNAL, NULL)) do { if (stonith_watchdog_timeout_ms <= 0) { crm_err("Ignoring watchdog fence device without " "stonith-watchdog-timeout set."); rv = -ENODEV; /* fall through to cleanup & return */ } else if (!pcmk__str_any_of(device->agent, STONITH_WATCHDOG_AGENT, STONITH_WATCHDOG_AGENT_INTERNAL, NULL)) { crm_err("Ignoring watchdog fence device with unknown " "agent '%s' unequal '" STONITH_WATCHDOG_AGENT "'.", device->agent?device->agent:""); rv = -ENODEV; /* fall through to cleanup & return */ } else if (!pcmk__str_eq(device->id, STONITH_WATCHDOG_ID, pcmk__str_none)) { crm_err("Ignoring watchdog fence device " "named %s !='"STONITH_WATCHDOG_ID"'.", device->id?device->id:""); rv = -ENODEV; /* fall through to cleanup & return */ } else { if (pcmk__str_eq(device->agent, STONITH_WATCHDOG_AGENT, pcmk__str_none)) { /* this either has an empty list or the targets configured for watchdog-fencing */ g_list_free_full(stonith_watchdog_targets, free); stonith_watchdog_targets = device->targets; device->targets = NULL; } if (node_does_watchdog_fencing(stonith_our_uname)) { g_list_free_full(device->targets, free); device->targets = stonith__parse_targets(stonith_our_uname); g_hash_table_replace(device->params, strdup(PCMK_STONITH_HOST_LIST), strdup(stonith_our_uname)); /* proceed as with any other stonith-device */ break; } crm_debug("Skip registration of watchdog fence device on node not in host-list."); /* cleanup and fall through to more cleanup and return */ device->targets = NULL; stonith_device_remove(device->id, from_cib); } free_device(device); return rv; } while (0); dup = device_has_duplicate(device); if (dup) { ndevices = g_hash_table_size(device_list); crm_debug("Device '%s' already in device list (%d active device%s)", device->id, ndevices, pcmk__plural_s(ndevices)); free_device(device); device = dup; dup = g_hash_table_lookup(device_list, device->id); dup->dirty = FALSE; } else { stonith_device_t *old = g_hash_table_lookup(device_list, device->id); if (from_cib && old && old->api_registered) { /* If the cib is writing over an entry that is shared with a stonith client, * copy any pending ops that currently exist on the old entry to the new one. * Otherwise the pending ops will be reported as failures */ crm_info("Overwriting existing entry for %s from CIB", device->id); device->pending_ops = old->pending_ops; device->api_registered = TRUE; old->pending_ops = NULL; if (device->pending_ops) { mainloop_set_trigger(device->work); } } g_hash_table_replace(device_list, device->id, device); ndevices = g_hash_table_size(device_list); crm_notice("Added '%s' to device list (%d active device%s)", device->id, ndevices, pcmk__plural_s(ndevices)); } if (desc) { *desc = device->id; } if (from_cib) { device->cib_registered = TRUE; } else { device->api_registered = TRUE; } return pcmk_ok; } void stonith_device_remove(const char *id, bool from_cib) { stonith_device_t *device = g_hash_table_lookup(device_list, id); guint ndevices = 0; if (!device) { ndevices = g_hash_table_size(device_list); crm_info("Device '%s' not found (%d active device%s)", id, ndevices, pcmk__plural_s(ndevices)); return; } if (from_cib) { device->cib_registered = FALSE; } else { device->verified = FALSE; device->api_registered = FALSE; } if (!device->cib_registered && !device->api_registered) { g_hash_table_remove(device_list, id); ndevices = g_hash_table_size(device_list); crm_info("Removed '%s' from device list (%d active device%s)", id, ndevices, pcmk__plural_s(ndevices)); } else { crm_trace("Not removing '%s' from device list (%d active) because " "still registered via:%s%s", id, g_hash_table_size(device_list), (device->cib_registered? " cib" : ""), (device->api_registered? " api" : "")); } } /*! * \internal * \brief Return the number of stonith levels registered for a node * * \param[in] tp Node's topology table entry * * \return Number of non-NULL levels in topology entry * \note This function is used only for log messages. */ static int count_active_levels(stonith_topology_t * tp) { int lpc = 0; int count = 0; for (lpc = 0; lpc < ST_LEVEL_MAX; lpc++) { if (tp->levels[lpc] != NULL) { count++; } } return count; } static void free_topology_entry(gpointer data) { stonith_topology_t *tp = data; int lpc = 0; for (lpc = 0; lpc < ST_LEVEL_MAX; lpc++) { if (tp->levels[lpc] != NULL) { g_list_free_full(tp->levels[lpc], free); } } free(tp->target); free(tp->target_value); free(tp->target_pattern); free(tp->target_attribute); free(tp); } void free_topology_list(void) { if (topology != NULL) { g_hash_table_destroy(topology); topology = NULL; } } void init_topology_list(void) { if (topology == NULL) { topology = pcmk__strkey_table(NULL, free_topology_entry); } } char *stonith_level_key(xmlNode *level, int mode) { if(mode == -1) { mode = stonith_level_kind(level); } switch(mode) { case 0: return crm_element_value_copy(level, XML_ATTR_STONITH_TARGET); case 1: return crm_element_value_copy(level, XML_ATTR_STONITH_TARGET_PATTERN); case 2: { const char *name = crm_element_value(level, XML_ATTR_STONITH_TARGET_ATTRIBUTE); const char *value = crm_element_value(level, XML_ATTR_STONITH_TARGET_VALUE); if(name && value) { return crm_strdup_printf("%s=%s", name, value); } } default: return crm_strdup_printf("Unknown-%d-%s", mode, ID(level)); } } int stonith_level_kind(xmlNode * level) { int mode = 0; const char *target = crm_element_value(level, XML_ATTR_STONITH_TARGET); if(target == NULL) { mode++; target = crm_element_value(level, XML_ATTR_STONITH_TARGET_PATTERN); } if(stand_alone == FALSE && target == NULL) { mode++; if(crm_element_value(level, XML_ATTR_STONITH_TARGET_ATTRIBUTE) == NULL) { mode++; } else if(crm_element_value(level, XML_ATTR_STONITH_TARGET_VALUE) == NULL) { mode++; } } return mode; } static stonith_key_value_t * parse_device_list(const char *devices) { int lpc = 0; int max = 0; int last = 0; stonith_key_value_t *output = NULL; if (devices == NULL) { return output; } max = strlen(devices); for (lpc = 0; lpc <= max; lpc++) { if (devices[lpc] == ',' || devices[lpc] == 0) { char *line = strndup(devices + last, lpc - last); output = stonith_key_value_add(output, NULL, line); free(line); last = lpc + 1; } } return output; } /*! * \internal - * \brief Register a STONITH level for a target + * \brief Register a fencing topology level for a target * * Given an XML request specifying the target name, level index, and device IDs * for the level, this will create an entry for the target in the global topology * table if one does not already exist, then append the specified device IDs to * the entry's device list for the specified level. * - * \param[in] msg XML request for STONITH level registration - * \param[out] desc If not NULL, will be set to string representation ("TARGET[LEVEL]") - * - * \return pcmk_ok on success, -EINVAL if XML does not specify valid level index + * \param[in] msg XML request for STONITH level registration + * \param[out] desc If not NULL, set to string representation "TARGET[LEVEL]" + * \param[out] result Where to set result of registration */ -int -stonith_level_register(xmlNode *msg, char **desc) +void +fenced_register_level(xmlNode *msg, char **desc, pcmk__action_result_t *result) { int id = 0; xmlNode *level; int mode; char *target; stonith_topology_t *tp; stonith_key_value_t *dIter = NULL; stonith_key_value_t *devices = NULL; + CRM_CHECK(result != NULL, return); + + if (msg == NULL) { + fenced_set_protocol_error(result); + return; + } + /* Allow the XML here to point to the level tag directly, or wrapped in * another tag. If directly, don't search by xpath, because it might give * multiple hits (e.g. if the XML is the CIB). */ if (pcmk__str_eq(TYPE(msg), XML_TAG_FENCING_LEVEL, pcmk__str_casei)) { level = msg; } else { - level = get_xpath_object("//" XML_TAG_FENCING_LEVEL, msg, LOG_ERR); + level = get_xpath_object("//" XML_TAG_FENCING_LEVEL, msg, LOG_WARNING); + } + if (level == NULL) { + fenced_set_protocol_error(result); + return; } - CRM_CHECK(level != NULL, return -EINVAL); mode = stonith_level_kind(level); + target = stonith_level_key(level, mode); crm_element_value_int(level, XML_ATTR_STONITH_INDEX, &id); if (desc) { *desc = crm_strdup_printf("%s[%d]", target, id); } - /* Sanity-check arguments */ - if (mode >= 3 || (id <= 0) || (id >= ST_LEVEL_MAX)) { - crm_trace("Could not add %s[%d] (%d) to the topology (%d active entries)", target, id, mode, g_hash_table_size(topology)); + // Ensure level ID is in allowed range + if ((id <= 0) || (id >= ST_LEVEL_MAX)) { + crm_warn("Ignoring topology registration for %s with invalid level %d", + target, id); free(target); - crm_log_xml_err(level, "Bad topology"); - return -EINVAL; + crm_log_xml_warn(level, "Bad level"); + pcmk__set_result(result, CRM_EX_INVALID_PARAM, PCMK_EXEC_INVALID, + "Invalid topology level"); + return; } /* Find or create topology table entry */ tp = g_hash_table_lookup(topology, target); if (tp == NULL) { tp = calloc(1, sizeof(stonith_topology_t)); + if (tp == NULL) { + pcmk__set_result(result, CRM_EX_ERROR, PCMK_EXEC_ERROR, + strerror(ENOMEM)); + return; + } tp->kind = mode; tp->target = target; tp->target_value = crm_element_value_copy(level, XML_ATTR_STONITH_TARGET_VALUE); tp->target_pattern = crm_element_value_copy(level, XML_ATTR_STONITH_TARGET_PATTERN); tp->target_attribute = crm_element_value_copy(level, XML_ATTR_STONITH_TARGET_ATTRIBUTE); g_hash_table_replace(topology, tp->target, tp); crm_trace("Added %s (%d) to the topology (%d active entries)", target, mode, g_hash_table_size(topology)); } else { free(target); } if (tp->levels[id] != NULL) { crm_info("Adding to the existing %s[%d] topology entry", tp->target, id); } devices = parse_device_list(crm_element_value(level, XML_ATTR_STONITH_DEVICES)); for (dIter = devices; dIter; dIter = dIter->next) { const char *device = dIter->value; crm_trace("Adding device '%s' for %s[%d]", device, tp->target, id); tp->levels[id] = g_list_append(tp->levels[id], strdup(device)); } stonith_key_value_freeall(devices, 1, 1); { int nlevels = count_active_levels(tp); crm_info("Target %s has %d active fencing level%s", tp->target, nlevels, pcmk__plural_s(nlevels)); } - return pcmk_ok; + + pcmk__set_result(result, CRM_EX_OK, PCMK_EXEC_DONE, NULL); } int stonith_level_remove(xmlNode *msg, char **desc) { int id = -1; stonith_topology_t *tp; char *target; /* Unlike additions, removal requests should always have one level tag */ xmlNode *level = get_xpath_object("//" XML_TAG_FENCING_LEVEL, msg, LOG_ERR); CRM_CHECK(level != NULL, return -EPROTO); target = stonith_level_key(level, -1); crm_element_value_int(level, XML_ATTR_STONITH_INDEX, &id); CRM_CHECK((id >= 0) && (id < ST_LEVEL_MAX), crm_log_xml_warn(msg, "invalid level"); free(target); return -EPROTO); if (desc) { *desc = crm_strdup_printf("%s[%d]", target, id); } tp = g_hash_table_lookup(topology, target); if (tp == NULL) { guint nentries = g_hash_table_size(topology); crm_info("No fencing topology found for %s (%d active %s)", target, nentries, pcmk__plural_alt(nentries, "entry", "entries")); } else if (id == 0 && g_hash_table_remove(topology, target)) { guint nentries = g_hash_table_size(topology); crm_info("Removed all fencing topology entries related to %s " "(%d active %s remaining)", target, nentries, pcmk__plural_alt(nentries, "entry", "entries")); } else if (tp->levels[id] != NULL) { guint nlevels; g_list_free_full(tp->levels[id], free); tp->levels[id] = NULL; nlevels = count_active_levels(tp); crm_info("Removed level %d from fencing topology for %s " "(%d active level%s remaining)", id, target, nlevels, pcmk__plural_s(nlevels)); } free(target); return pcmk_ok; } static char * list_to_string(GList *list, const char *delim, gboolean terminate_with_delim) { int max = g_list_length(list); size_t delim_len = delim?strlen(delim):0; size_t alloc_size = 1 + (max?((max-1+(terminate_with_delim?1:0))*delim_len):0); char *rv; GList *gIter; for (gIter = list; gIter != NULL; gIter = gIter->next) { const char *value = (const char *) gIter->data; alloc_size += strlen(value); } rv = calloc(alloc_size, sizeof(char)); if (rv) { char *pos = rv; const char *lead_delim = ""; for (gIter = list; gIter != NULL; gIter = gIter->next) { const char *value = (const char *) gIter->data; pos = &pos[sprintf(pos, "%s%s", lead_delim, value)]; lead_delim = delim; } if (max && terminate_with_delim) { sprintf(pos, "%s", delim); } } return rv; } /*! * \internal * \brief Execute a fence agent action directly (and asynchronously) * * Handle a STONITH_OP_EXEC API message by scheduling a requested agent action * directly on a specified device. Only list, monitor, and status actions are * expected to use this call, though it should work with any agent command. * * \param[in] msg Request XML specifying action * \param[out] result Where to store result of action * * \note If the action is monitor, the device must be registered via the API * (CIB registration is not sufficient), because monitor should not be * possible unless the device is "started" (API registered). */ static void execute_agent_action(xmlNode *msg, pcmk__action_result_t *result) { xmlNode *dev = get_xpath_object("//" F_STONITH_DEVICE, msg, LOG_ERR); xmlNode *op = get_xpath_object("//@" F_STONITH_ACTION, msg, LOG_ERR); const char *id = crm_element_value(dev, F_STONITH_DEVICE); const char *action = crm_element_value(op, F_STONITH_ACTION); async_command_t *cmd = NULL; stonith_device_t *device = NULL; if ((id == NULL) || (action == NULL)) { crm_info("Malformed API action request: device %s, action %s", (id? id : "not specified"), (action? action : "not specified")); fenced_set_protocol_error(result); return; } if (pcmk__str_eq(id, STONITH_WATCHDOG_ID, pcmk__str_none)) { // Watchdog agent actions are implemented internally if (stonith_watchdog_timeout_ms <= 0) { pcmk__set_result(result, CRM_EX_ERROR, PCMK_EXEC_NO_FENCE_DEVICE, "Watchdog fence device not configured"); return; } else if (pcmk__str_eq(action, "list", pcmk__str_casei)) { pcmk__set_result(result, CRM_EX_OK, PCMK_EXEC_DONE, NULL); pcmk__set_result_output(result, list_to_string(stonith_watchdog_targets, "\n", TRUE), NULL); return; } else if (pcmk__str_eq(action, "monitor", pcmk__str_casei)) { pcmk__set_result(result, CRM_EX_OK, PCMK_EXEC_DONE, NULL); return; } } device = g_hash_table_lookup(device_list, id); if (device == NULL) { crm_info("Ignoring API '%s' action request because device %s not found", action, id); pcmk__set_result(result, CRM_EX_ERROR, PCMK_EXEC_NO_FENCE_DEVICE, NULL); return; } else if (!device->api_registered && !strcmp(action, "monitor")) { // Monitors may run only on "started" (API-registered) devices crm_info("Ignoring API '%s' action request because device %s not active", action, id); pcmk__set_result(result, CRM_EX_ERROR, PCMK_EXEC_NO_FENCE_DEVICE, "Fence device not active"); return; } cmd = create_async_command(msg); if (cmd == NULL) { fenced_set_protocol_error(result); return; } schedule_stonith_command(cmd, device); pcmk__set_result(result, CRM_EX_OK, PCMK_EXEC_PENDING, NULL); } static void search_devices_record_result(struct device_search_s *search, const char *device, gboolean can_fence) { search->replies_received++; if (can_fence && device) { search->capable = g_list_append(search->capable, strdup(device)); } if (search->replies_needed == search->replies_received) { guint ndevices = g_list_length(search->capable); crm_debug("Search found %d device%s that can perform '%s' targeting %s", ndevices, pcmk__plural_s(ndevices), (search->action? search->action : "unknown action"), (search->host? search->host : "any node")); search->callback(search->capable, search->user_data); free(search->host); free(search->action); free(search); } } /*! * \internal * \brief Check whether the local host is allowed to execute a fencing action * * \param[in] device Fence device to check * \param[in] action Fence action to check * \param[in] target Hostname of fence target * \param[in] allow_suicide Whether self-fencing is allowed for this operation * * \return TRUE if local host is allowed to execute action, FALSE otherwise */ static gboolean localhost_is_eligible(const stonith_device_t *device, const char *action, const char *target, gboolean allow_suicide) { gboolean localhost_is_target = pcmk__str_eq(target, stonith_our_uname, pcmk__str_casei); if (device && action && device->on_target_actions && strstr(device->on_target_actions, action)) { if (!localhost_is_target) { crm_trace("Operation '%s' using %s can only be executed for " "local host, not %s", action, device->id, target); return FALSE; } } else if (localhost_is_target && !allow_suicide) { crm_trace("'%s' operation does not support self-fencing", action); return FALSE; } return TRUE; } static void can_fence_host_with_device(stonith_device_t * dev, struct device_search_s *search) { gboolean can = FALSE; const char *check_type = NULL; const char *host = search->host; const char *alias = NULL; CRM_LOG_ASSERT(dev != NULL); if (dev == NULL) { goto search_report_results; } else if (host == NULL) { can = TRUE; goto search_report_results; } /* Short-circuit query if this host is not allowed to perform the action */ if (pcmk__str_eq(search->action, "reboot", pcmk__str_casei)) { /* A "reboot" *might* get remapped to "off" then "on", so short-circuit * only if all three are disallowed. If only one or two are disallowed, * we'll report that with the results. We never allow suicide for * remapped "on" operations because the host is off at that point. */ if (!localhost_is_eligible(dev, "reboot", host, search->allow_suicide) && !localhost_is_eligible(dev, "off", host, search->allow_suicide) && !localhost_is_eligible(dev, "on", host, FALSE)) { goto search_report_results; } } else if (!localhost_is_eligible(dev, search->action, host, search->allow_suicide)) { goto search_report_results; } alias = g_hash_table_lookup(dev->aliases, host); if (alias == NULL) { alias = host; } check_type = target_list_type(dev); if (pcmk__str_eq(check_type, "none", pcmk__str_casei)) { can = TRUE; } else if (pcmk__str_eq(check_type, "static-list", pcmk__str_casei)) { /* Presence in the hostmap is sufficient * Only use if all hosts on which the device can be active can always fence all listed hosts */ if (pcmk__str_in_list(host, dev->targets, pcmk__str_casei)) { can = TRUE; } else if (g_hash_table_lookup(dev->params, PCMK_STONITH_HOST_MAP) && g_hash_table_lookup(dev->aliases, host)) { can = TRUE; } } else if (pcmk__str_eq(check_type, "dynamic-list", pcmk__str_casei)) { time_t now = time(NULL); if (dev->targets == NULL || dev->targets_age + 60 < now) { int device_timeout = get_action_timeout(dev, "list", search->per_device_timeout); if (device_timeout > search->per_device_timeout) { crm_notice("Since the pcmk_list_timeout(%ds) parameter of %s is larger than stonith-timeout(%ds), timeout may occur", device_timeout, dev->id, search->per_device_timeout); } crm_trace("Running '%s' to check whether %s is eligible to fence %s (%s)", check_type, dev->id, search->host, search->action); schedule_internal_command(__func__, dev, "list", NULL, search->per_device_timeout, search, dynamic_list_search_cb); /* we'll respond to this search request async in the cb */ return; } if (pcmk__str_in_list(alias, dev->targets, pcmk__str_casei)) { can = TRUE; } } else if (pcmk__str_eq(check_type, "status", pcmk__str_casei)) { int device_timeout = get_action_timeout(dev, check_type, search->per_device_timeout); if (device_timeout > search->per_device_timeout) { crm_notice("Since the pcmk_status_timeout(%ds) parameter of %s is larger than stonith-timeout(%ds), timeout may occur", device_timeout, dev->id, search->per_device_timeout); } crm_trace("Running '%s' to check whether %s is eligible to fence %s (%s)", check_type, dev->id, search->host, search->action); schedule_internal_command(__func__, dev, "status", search->host, search->per_device_timeout, search, status_search_cb); /* we'll respond to this search request async in the cb */ return; } else { crm_err("Invalid value for " PCMK_STONITH_HOST_CHECK ": %s", check_type); check_type = "Invalid " PCMK_STONITH_HOST_CHECK; } if (pcmk__str_eq(host, alias, pcmk__str_casei)) { crm_notice("%s is%s eligible to fence (%s) %s: %s", dev->id, (can? "" : " not"), search->action, host, check_type); } else { crm_notice("%s is%s eligible to fence (%s) %s (aka. '%s'): %s", dev->id, (can? "" : " not"), search->action, host, alias, check_type); } search_report_results: search_devices_record_result(search, dev ? dev->id : NULL, can); } static void search_devices(gpointer key, gpointer value, gpointer user_data) { stonith_device_t *dev = value; struct device_search_s *search = user_data; can_fence_host_with_device(dev, search); } #define DEFAULT_QUERY_TIMEOUT 20 static void get_capable_devices(const char *host, const char *action, int timeout, bool suicide, void *user_data, void (*callback) (GList * devices, void *user_data)) { struct device_search_s *search; guint ndevices = g_hash_table_size(device_list); if (ndevices == 0) { callback(NULL, user_data); return; } search = calloc(1, sizeof(struct device_search_s)); if (!search) { crm_crit("Cannot search for capable fence devices: %s", strerror(ENOMEM)); callback(NULL, user_data); return; } search->host = host ? strdup(host) : NULL; search->action = action ? strdup(action) : NULL; search->per_device_timeout = timeout; search->allow_suicide = suicide; search->callback = callback; search->user_data = user_data; /* We are guaranteed this many replies, even if a device is * unregistered while the search is in progress. */ search->replies_needed = ndevices; crm_debug("Searching %d device%s to see which can execute '%s' targeting %s", ndevices, pcmk__plural_s(ndevices), (search->action? search->action : "unknown action"), (search->host? search->host : "any node")); g_hash_table_foreach(device_list, search_devices, search); } struct st_query_data { xmlNode *reply; char *remote_peer; char *client_id; char *target; char *action; int call_options; }; /*! * \internal * \brief Add action-specific attributes to query reply XML * * \param[in,out] xml XML to add attributes to * \param[in] action Fence action * \param[in] device Fence device * \param[in] target Fence target */ static void add_action_specific_attributes(xmlNode *xml, const char *action, stonith_device_t *device, const char *target) { int action_specific_timeout; int delay_max; int delay_base; CRM_CHECK(xml && action && device, return); if (is_action_required(action, device)) { crm_trace("Action '%s' is required using %s", action, device->id); crm_xml_add_int(xml, F_STONITH_DEVICE_REQUIRED, 1); } action_specific_timeout = get_action_timeout(device, action, 0); if (action_specific_timeout) { crm_trace("Action '%s' has timeout %dms using %s", action, action_specific_timeout, device->id); crm_xml_add_int(xml, F_STONITH_ACTION_TIMEOUT, action_specific_timeout); } delay_max = get_action_delay_max(device, action); if (delay_max > 0) { crm_trace("Action '%s' has maximum random delay %dms using %s", action, delay_max, device->id); crm_xml_add_int(xml, F_STONITH_DELAY_MAX, delay_max / 1000); } delay_base = get_action_delay_base(device, action, target); if (delay_base > 0) { crm_xml_add_int(xml, F_STONITH_DELAY_BASE, delay_base / 1000); } if ((delay_max > 0) && (delay_base == 0)) { crm_trace("Action '%s' has maximum random delay %dms using %s", action, delay_max, device->id); } else if ((delay_max == 0) && (delay_base > 0)) { crm_trace("Action '%s' has a static delay of %dms using %s", action, delay_base, device->id); } else if ((delay_max > 0) && (delay_base > 0)) { crm_trace("Action '%s' has a minimum delay of %dms and a randomly chosen " "maximum delay of %dms using %s", action, delay_base, delay_max, device->id); } } /*! * \internal * \brief Add "disallowed" attribute to query reply XML if appropriate * * \param[in,out] xml XML to add attribute to * \param[in] action Fence action * \param[in] device Fence device * \param[in] target Fence target * \param[in] allow_suicide Whether self-fencing is allowed */ static void add_disallowed(xmlNode *xml, const char *action, stonith_device_t *device, const char *target, gboolean allow_suicide) { if (!localhost_is_eligible(device, action, target, allow_suicide)) { crm_trace("Action '%s' using %s is disallowed for local host", action, device->id); pcmk__xe_set_bool_attr(xml, F_STONITH_ACTION_DISALLOWED, true); } } /*! * \internal * \brief Add child element with action-specific values to query reply XML * * \param[in,out] xml XML to add attribute to * \param[in] action Fence action * \param[in] device Fence device * \param[in] target Fence target * \param[in] allow_suicide Whether self-fencing is allowed */ static void add_action_reply(xmlNode *xml, const char *action, stonith_device_t *device, const char *target, gboolean allow_suicide) { xmlNode *child = create_xml_node(xml, F_STONITH_ACTION); crm_xml_add(child, XML_ATTR_ID, action); add_action_specific_attributes(child, action, device, target); add_disallowed(child, action, device, target, allow_suicide); } static void stonith_query_capable_device_cb(GList * devices, void *user_data) { struct st_query_data *query = user_data; int available_devices = 0; xmlNode *dev = NULL; xmlNode *list = NULL; GList *lpc = NULL; /* Pack the results into XML */ list = create_xml_node(NULL, __func__); crm_xml_add(list, F_STONITH_TARGET, query->target); for (lpc = devices; lpc != NULL; lpc = lpc->next) { stonith_device_t *device = g_hash_table_lookup(device_list, lpc->data); const char *action = query->action; if (!device) { /* It is possible the device got unregistered while * determining who can fence the target */ continue; } available_devices++; dev = create_xml_node(list, F_STONITH_DEVICE); crm_xml_add(dev, XML_ATTR_ID, device->id); crm_xml_add(dev, "namespace", device->namespace); crm_xml_add(dev, "agent", device->agent); crm_xml_add_int(dev, F_STONITH_DEVICE_VERIFIED, device->verified); /* If the originating fencer wants to reboot the node, and we have a * capable device that doesn't support "reboot", remap to "off" instead. */ if (!pcmk_is_set(device->flags, st_device_supports_reboot) && pcmk__str_eq(query->action, "reboot", pcmk__str_casei)) { crm_trace("%s doesn't support reboot, using values for off instead", device->id); action = "off"; } /* Add action-specific values if available */ add_action_specific_attributes(dev, action, device, query->target); if (pcmk__str_eq(query->action, "reboot", pcmk__str_casei)) { /* A "reboot" *might* get remapped to "off" then "on", so after * sending the "reboot"-specific values in the main element, we add * sub-elements for "off" and "on" values. * * We short-circuited earlier if "reboot", "off" and "on" are all * disallowed for the local host. However if only one or two are * disallowed, we send back the results and mark which ones are * disallowed. If "reboot" is disallowed, this might cause problems * with older fencer versions, which won't check for it. Older * versions will ignore "off" and "on", so they are not a problem. */ add_disallowed(dev, action, device, query->target, pcmk_is_set(query->call_options, st_opt_allow_suicide)); add_action_reply(dev, "off", device, query->target, pcmk_is_set(query->call_options, st_opt_allow_suicide)); add_action_reply(dev, "on", device, query->target, FALSE); } /* A query without a target wants device parameters */ if (query->target == NULL) { xmlNode *attrs = create_xml_node(dev, XML_TAG_ATTRS); g_hash_table_foreach(device->params, hash2field, attrs); } } crm_xml_add_int(list, F_STONITH_AVAILABLE_DEVICES, available_devices); if (query->target) { crm_debug("Found %d matching device%s for target '%s'", available_devices, pcmk__plural_s(available_devices), query->target); } else { crm_debug("%d device%s installed", available_devices, pcmk__plural_s(available_devices)); } if (list != NULL) { crm_log_xml_trace(list, "Add query results"); add_message_xml(query->reply, F_STONITH_CALLDATA, list); } stonith_send_reply(query->reply, query->call_options, query->remote_peer, query->client_id); free_xml(query->reply); free(query->remote_peer); free(query->client_id); free(query->target); free(query->action); free(query); free_xml(list); g_list_free_full(devices, free); } static void stonith_query(xmlNode * msg, const char *remote_peer, const char *client_id, int call_options) { struct st_query_data *query = NULL; const char *action = NULL; const char *target = NULL; int timeout = 0; xmlNode *dev = get_xpath_object("//@" F_STONITH_ACTION, msg, LOG_NEVER); crm_element_value_int(msg, F_STONITH_TIMEOUT, &timeout); if (dev) { const char *device = crm_element_value(dev, F_STONITH_DEVICE); target = crm_element_value(dev, F_STONITH_TARGET); action = crm_element_value(dev, F_STONITH_ACTION); if (device && pcmk__str_eq(device, "manual_ack", pcmk__str_casei)) { /* No query or reply necessary */ return; } } crm_log_xml_debug(msg, "Query"); query = calloc(1, sizeof(struct st_query_data)); query->reply = stonith_construct_reply(msg, NULL, NULL, pcmk_ok); query->remote_peer = remote_peer ? strdup(remote_peer) : NULL; query->client_id = client_id ? strdup(client_id) : NULL; query->target = target ? strdup(target) : NULL; query->action = action ? strdup(action) : NULL; query->call_options = call_options; get_capable_devices(target, action, timeout, pcmk_is_set(call_options, st_opt_allow_suicide), query, stonith_query_capable_device_cb); } /*! * \internal * \brief Log the result of an asynchronous command * * \param[in] cmd Command the result is for * \param[in] result Result of command * \param[in] pid Process ID of command, if available * \param[in] next Alternate device that will be tried if command failed * \param[in] op_merged Whether this command was merged with an earlier one */ static void log_async_result(async_command_t *cmd, const pcmk__action_result_t *result, int pid, const char *next, bool op_merged) { int log_level = LOG_ERR; int output_log_level = LOG_NEVER; guint devices_remaining = g_list_length(cmd->device_next); GString *msg = g_string_sized_new(80); // Reasonable starting size // Choose log levels appropriately if we have a result if (pcmk__result_ok(result)) { log_level = (cmd->victim == NULL)? LOG_DEBUG : LOG_NOTICE; if ((result->action_stdout != NULL) && !pcmk__str_eq(cmd->action, "metadata", pcmk__str_casei)) { output_log_level = LOG_DEBUG; } next = NULL; } else { log_level = (cmd->victim == NULL)? LOG_NOTICE : LOG_ERR; if ((result->action_stdout != NULL) && !pcmk__str_eq(cmd->action, "metadata", pcmk__str_casei)) { output_log_level = LOG_WARNING; } } // Build the log message piece by piece g_string_printf(msg, "Operation '%s' ", cmd->action); if (pid != 0) { g_string_append_printf(msg, "[%d] ", pid); } if (cmd->victim != NULL) { g_string_append_printf(msg, "targeting %s ", cmd->victim); } g_string_append_printf(msg, "using %s ", cmd->device); // Add exit status or execution status as appropriate if (result->execution_status == PCMK_EXEC_DONE) { g_string_append_printf(msg, "returned %d", result->exit_status); } else { g_string_append_printf(msg, "could not be executed: %s", pcmk_exec_status_str(result->execution_status)); } // Add exit reason and next device if appropriate if (result->exit_reason != NULL) { g_string_append_printf(msg, " (%s)", result->exit_reason); } if (next != NULL) { g_string_append_printf(msg, ", retrying with %s", next); } if (devices_remaining > 0) { g_string_append_printf(msg, " (%u device%s remaining)", (unsigned int) devices_remaining, pcmk__plural_s(devices_remaining)); } g_string_append_printf(msg, " " CRM_XS " %scall %d from %s", (op_merged? "merged " : ""), cmd->id, cmd->client_name); // Log the result do_crm_log(log_level, "%s", msg->str); g_string_free(msg, TRUE); // Log the output (which may have multiple lines), if appropriate if (output_log_level != LOG_NEVER) { char *prefix = crm_strdup_printf("%s[%d]", cmd->device, pid); crm_log_output(output_log_level, prefix, result->action_stdout); free(prefix); } } /*! * \internal * \brief Reply to requester after asynchronous command completion * * \param[in] cmd Command that completed * \param[in] result Result of command * \param[in] pid Process ID of command, if available * \param[in] merged If true, command was merged with another, not executed */ static void send_async_reply(async_command_t *cmd, const pcmk__action_result_t *result, int pid, bool merged) { xmlNode *reply = NULL; CRM_CHECK((cmd != NULL) && (result != NULL), return); log_async_result(cmd, result, pid, NULL, merged); reply = construct_async_reply(cmd, result); if (merged) { pcmk__xe_set_bool_attr(reply, F_STONITH_MERGED, true); } if (!stand_alone && pcmk__is_fencing_action(cmd->action) && pcmk__str_eq(cmd->origin, cmd->victim, pcmk__str_casei)) { /* The target was also the originator, so broadcast the result on its * behalf (since it will be unable to). */ crm_trace("Broadcast '%s' result for %s (target was also originator)", cmd->action, cmd->victim); crm_xml_add(reply, F_SUBTYPE, "broadcast"); crm_xml_add(reply, F_STONITH_OPERATION, T_STONITH_NOTIFY); send_cluster_message(NULL, crm_msg_stonith_ng, reply, FALSE); } else { // Reply only to the originator stonith_send_reply(reply, cmd->options, cmd->origin, cmd->client); } crm_log_xml_trace(reply, "Reply"); free_xml(reply); if (stand_alone) { /* Do notification with a clean data object */ xmlNode *notify_data = create_xml_node(NULL, T_STONITH_NOTIFY_FENCE); stonith__xe_set_result(notify_data, result); crm_xml_add(notify_data, F_STONITH_TARGET, cmd->victim); crm_xml_add(notify_data, F_STONITH_OPERATION, cmd->op); crm_xml_add(notify_data, F_STONITH_DELEGATE, "localhost"); crm_xml_add(notify_data, F_STONITH_DEVICE, cmd->device); crm_xml_add(notify_data, F_STONITH_REMOTE_OP_ID, cmd->remote_op_id); crm_xml_add(notify_data, F_STONITH_ORIGIN, cmd->client); do_stonith_notify(T_STONITH_NOTIFY_FENCE, pcmk_rc2legacy(stonith__result2rc(result)), notify_data); do_stonith_notify(T_STONITH_NOTIFY_HISTORY, pcmk_ok, NULL); } } static void cancel_stonith_command(async_command_t * cmd) { stonith_device_t *device; CRM_CHECK(cmd != NULL, return); if (!cmd->device) { return; } device = g_hash_table_lookup(device_list, cmd->device); if (device) { crm_trace("Cancel scheduled '%s' action using %s", cmd->action, device->id); device->pending_ops = g_list_remove(device->pending_ops, cmd); } } static void st_child_done(int pid, const pcmk__action_result_t *result, void *user_data) { stonith_device_t *device = NULL; stonith_device_t *next_device = NULL; async_command_t *cmd = user_data; GList *gIter = NULL; GList *gIterNext = NULL; CRM_CHECK(cmd != NULL, return); cmd->active_on = NULL; /* The device is ready to do something else now */ device = g_hash_table_lookup(device_list, cmd->device); if (device) { if (!device->verified && pcmk__result_ok(result) && (pcmk__strcase_any_of(cmd->action, "list", "monitor", "status", NULL))) { device->verified = TRUE; } mainloop_set_trigger(device->work); } if (pcmk__result_ok(result)) { GList *iter; /* see if there are any required devices left to execute for this op */ for (iter = cmd->device_next; iter != NULL; iter = iter->next) { next_device = g_hash_table_lookup(device_list, iter->data); if (next_device != NULL && is_action_required(cmd->action, next_device)) { cmd->device_next = iter->next; break; } next_device = NULL; } } else if ((cmd->device_next != NULL) && !is_action_required(cmd->action, device)) { /* if this device didn't work out, see if there are any others we can try. * if the failed device was 'required', we can't pick another device. */ next_device = g_hash_table_lookup(device_list, cmd->device_next->data); cmd->device_next = cmd->device_next->next; } /* this operation requires more fencing, hooray! */ if (next_device) { log_async_result(cmd, result, pid, next_device->id, false); schedule_stonith_command(cmd, next_device); /* Prevent cmd from being freed */ cmd = NULL; goto done; } send_async_reply(cmd, result, pid, false); if (!pcmk__result_ok(result)) { goto done; } /* Check to see if any operations are scheduled to do the exact * same thing that just completed. If so, rather than * performing the same fencing operation twice, return the result * of this operation for all pending commands it matches. */ for (gIter = cmd_list; gIter != NULL; gIter = gIterNext) { async_command_t *cmd_other = gIter->data; gIterNext = gIter->next; if (cmd == cmd_other) { continue; } /* A pending scheduled command matches the command that just finished if. * 1. The client connections are different. * 2. The node victim is the same. * 3. The fencing action is the same. * 4. The device scheduled to execute the action is the same. */ if (pcmk__str_eq(cmd->client, cmd_other->client, pcmk__str_casei) || !pcmk__str_eq(cmd->victim, cmd_other->victim, pcmk__str_casei) || !pcmk__str_eq(cmd->action, cmd_other->action, pcmk__str_casei) || !pcmk__str_eq(cmd->device, cmd_other->device, pcmk__str_casei)) { continue; } /* Duplicate merging will do the right thing for either type of remapped * reboot. If the executing fencer remapped an unsupported reboot to * off, then cmd->action will be reboot and will be merged with any * other reboot requests. If the originating fencer remapped a * topology reboot to off then on, we will get here once with * cmd->action "off" and once with "on", and they will be merged * separately with similar requests. */ crm_notice("Merging fencing action '%s' targeting %s originating from " "client %s with identical fencing request from client %s", cmd_other->action, cmd_other->victim, cmd_other->client_name, cmd->client_name); cmd_list = g_list_remove_link(cmd_list, gIter); send_async_reply(cmd_other, result, pid, true); cancel_stonith_command(cmd_other); free_async_command(cmd_other); g_list_free_1(gIter); } done: free_async_command(cmd); } static gint sort_device_priority(gconstpointer a, gconstpointer b) { const stonith_device_t *dev_a = a; const stonith_device_t *dev_b = b; if (dev_a->priority > dev_b->priority) { return -1; } else if (dev_a->priority < dev_b->priority) { return 1; } return 0; } static void stonith_fence_get_devices_cb(GList * devices, void *user_data) { async_command_t *cmd = user_data; stonith_device_t *device = NULL; guint ndevices = g_list_length(devices); crm_info("Found %d matching device%s for target '%s'", ndevices, pcmk__plural_s(ndevices), cmd->victim); if (devices != NULL) { /* Order based on priority */ devices = g_list_sort(devices, sort_device_priority); device = g_hash_table_lookup(device_list, devices->data); } if (device == NULL) { // No device found pcmk__action_result_t result = PCMK__UNKNOWN_RESULT; pcmk__set_result(&result, CRM_EX_ERROR, PCMK_EXEC_NO_FENCE_DEVICE, "No fence device configured for target"); send_async_reply(cmd, &result, 0, false); pcmk__reset_result(&result); free_async_command(cmd); g_list_free_full(devices, free); } else { // Device found, schedule it for fencing cmd->device_list = devices; cmd->device_next = devices->next; schedule_stonith_command(cmd, device); } } /*! * \internal * \brief Execute a fence action via the local node * * \param[in] msg Fencing request * \param[out] result Where to store result of fence action */ static void fence_locally(xmlNode *msg, pcmk__action_result_t *result) { const char *device_id = NULL; stonith_device_t *device = NULL; async_command_t *cmd = create_async_command(msg); xmlNode *dev = get_xpath_object("//@" F_STONITH_TARGET, msg, LOG_ERR); CRM_CHECK(result != NULL, return); if (cmd == NULL) { fenced_set_protocol_error(result); return; } device_id = crm_element_value(dev, F_STONITH_DEVICE); if (device_id != NULL) { device = g_hash_table_lookup(device_list, device_id); if (device == NULL) { crm_err("Requested device '%s' is not available", device_id); pcmk__set_result(result, CRM_EX_ERROR, PCMK_EXEC_NO_FENCE_DEVICE, "Requested fence device not found"); return; } schedule_stonith_command(cmd, device); } else { const char *host = crm_element_value(dev, F_STONITH_TARGET); if (pcmk_is_set(cmd->options, st_opt_cs_nodeid)) { int nodeid = 0; crm_node_t *node = NULL; pcmk__scan_min_int(host, &nodeid, 0); node = pcmk__search_known_node_cache(nodeid, NULL, CRM_GET_PEER_ANY); if (node != NULL) { host = node->uname; } } /* If we get to here, then self-fencing is implicitly allowed */ get_capable_devices(host, cmd->action, cmd->default_timeout, TRUE, cmd, stonith_fence_get_devices_cb); } pcmk__set_result(result, CRM_EX_OK, PCMK_EXEC_PENDING, NULL); } xmlNode * stonith_construct_reply(xmlNode * request, const char *output, xmlNode * data, int rc) { xmlNode *reply = NULL; reply = create_xml_node(NULL, T_STONITH_REPLY); crm_xml_add(reply, "st_origin", __func__); crm_xml_add(reply, F_TYPE, T_STONITH_NG); crm_xml_add(reply, F_STONITH_OUTPUT, output); crm_xml_add_int(reply, F_STONITH_RC, rc); if (request == NULL) { /* Most likely, this is the result of a stonith operation that was * initiated before we came up. Unfortunately that means we lack enough * information to provide clients with a full result. * * @TODO Maybe synchronize this information at start-up? */ crm_warn("Missing request information for client notifications for " "operation with result %d (initiated before we came up?)", rc); } else { const char *name = NULL; const char *value = NULL; const char *names[] = { F_STONITH_OPERATION, F_STONITH_CALLID, F_STONITH_CLIENTID, F_STONITH_CLIENTNAME, F_STONITH_REMOTE_OP_ID, F_STONITH_CALLOPTS }; crm_trace("Creating a result reply with%s reply output (rc=%d)", (data? "" : "out"), rc); for (int lpc = 0; lpc < PCMK__NELEM(names); lpc++) { name = names[lpc]; value = crm_element_value(request, name); crm_xml_add(reply, name, value); } if (data != NULL) { add_message_xml(reply, F_STONITH_CALLDATA, data); } } return reply; } /*! * \internal * \brief Build an XML reply to an asynchronous fencing command * * \param[in] cmd Fencing command that reply is for * \param[in] result Command result */ static xmlNode * construct_async_reply(async_command_t *cmd, const pcmk__action_result_t *result) { xmlNode *reply = create_xml_node(NULL, T_STONITH_REPLY); crm_xml_add(reply, "st_origin", __func__); crm_xml_add(reply, F_TYPE, T_STONITH_NG); crm_xml_add(reply, F_STONITH_OPERATION, cmd->op); crm_xml_add(reply, F_STONITH_DEVICE, cmd->device); crm_xml_add(reply, F_STONITH_REMOTE_OP_ID, cmd->remote_op_id); crm_xml_add(reply, F_STONITH_CLIENTID, cmd->client); crm_xml_add(reply, F_STONITH_CLIENTNAME, cmd->client_name); crm_xml_add(reply, F_STONITH_TARGET, cmd->victim); crm_xml_add(reply, F_STONITH_ACTION, cmd->op); crm_xml_add(reply, F_STONITH_ORIGIN, cmd->origin); crm_xml_add_int(reply, F_STONITH_CALLID, cmd->id); crm_xml_add_int(reply, F_STONITH_CALLOPTS, cmd->options); stonith__xe_set_result(reply, result); return reply; } bool fencing_peer_active(crm_node_t *peer) { if (peer == NULL) { return FALSE; } else if (peer->uname == NULL) { return FALSE; } else if (pcmk_is_set(peer->processes, crm_get_cluster_proc())) { return TRUE; } return FALSE; } void set_fencing_completed(remote_fencing_op_t * op) { #ifdef CLOCK_MONOTONIC struct timespec tv; clock_gettime(CLOCK_MONOTONIC, &tv); op->completed = tv.tv_sec; op->completed_nsec = tv.tv_nsec; #else op->completed = time(NULL); op->completed_nsec = 0L; #endif } /*! * \internal * \brief Look for alternate node needed if local node shouldn't fence target * * \param[in] target Node that must be fenced * * \return Name of an alternate node that should fence \p target if any, * or NULL otherwise */ static const char * check_alternate_host(const char *target) { const char *alternate_host = NULL; crm_trace("Checking if we (%s) can fence %s", stonith_our_uname, target); if (find_topology_for_host(target) && pcmk__str_eq(target, stonith_our_uname, pcmk__str_casei)) { GHashTableIter gIter; crm_node_t *entry = NULL; g_hash_table_iter_init(&gIter, crm_peer_cache); while (g_hash_table_iter_next(&gIter, NULL, (void **)&entry)) { crm_trace("Checking for %s.%d != %s", entry->uname, entry->id, target); if (fencing_peer_active(entry) && !pcmk__str_eq(entry->uname, target, pcmk__str_casei)) { alternate_host = entry->uname; break; } } if (alternate_host == NULL) { crm_err("No alternate host available to handle request " "for self-fencing with topology"); g_hash_table_iter_init(&gIter, crm_peer_cache); while (g_hash_table_iter_next(&gIter, NULL, (void **)&entry)) { crm_notice("Peer[%d] %s", entry->id, entry->uname); } } } return alternate_host; } /*! * \internal * \brief Send a reply to a CPG peer or IPC client * * \param[in] reply XML reply to send * \param[in] call_options Send synchronously if st_opt_sync_call is set here * \param[in] remote_peer If not NULL, name of peer node to send CPG reply * \param[in] client_id If not NULL, name of client to send IPC reply */ static void stonith_send_reply(xmlNode *reply, int call_options, const char *remote_peer, const char *client_id) { CRM_CHECK((reply != NULL) && ((remote_peer != NULL) || (client_id != NULL)), return); if (remote_peer == NULL) { do_local_reply(reply, client_id, pcmk_is_set(call_options, st_opt_sync_call), FALSE); } else { send_cluster_message(crm_get_peer(0, remote_peer), crm_msg_stonith_ng, reply, FALSE); } } static void remove_relay_op(xmlNode * request) { xmlNode *dev = get_xpath_object("//@" F_STONITH_ACTION, request, LOG_TRACE); const char *relay_op_id = NULL; const char *op_id = NULL; const char *client_name = NULL; const char *target = NULL; remote_fencing_op_t *relay_op = NULL; if (dev) { target = crm_element_value(dev, F_STONITH_TARGET); } relay_op_id = crm_element_value(request, F_STONITH_REMOTE_OP_ID_RELAY); op_id = crm_element_value(request, F_STONITH_REMOTE_OP_ID); client_name = crm_element_value(request, F_STONITH_CLIENTNAME); /* Delete RELAY operation. */ if (relay_op_id && target && pcmk__str_eq(target, stonith_our_uname, pcmk__str_casei)) { relay_op = g_hash_table_lookup(stonith_remote_op_list, relay_op_id); if (relay_op) { GHashTableIter iter; remote_fencing_op_t *list_op = NULL; g_hash_table_iter_init(&iter, stonith_remote_op_list); /* If the operation to be deleted is registered as a duplicate, delete the registration. */ while (g_hash_table_iter_next(&iter, NULL, (void **)&list_op)) { GList *dup_iter = NULL; if (list_op != relay_op) { for (dup_iter = list_op->duplicates; dup_iter != NULL; dup_iter = dup_iter->next) { remote_fencing_op_t *other = dup_iter->data; if (other == relay_op) { other->duplicates = g_list_remove(other->duplicates, relay_op); break; } } } } crm_debug("Deleting relay op %s ('%s' targeting %s for %s), " "replaced by op %s ('%s' targeting %s for %s)", relay_op->id, relay_op->action, relay_op->target, relay_op->client_name, op_id, relay_op->action, target, client_name); g_hash_table_remove(stonith_remote_op_list, relay_op_id); } } } /*! * \internal * \brief Check whether an API request was sent by a privileged user * * API commands related to fencing configuration may be done only by privileged * IPC users (i.e. root or hacluster), because all other users should go through * the CIB to have ACLs applied. If no client was given, this is a peer request, * which is always allowed. * * \param[in] c IPC client that sent request (or NULL if sent by CPG peer) * \param[in] op Requested API operation (for logging only) * * \return true if sender is peer or privileged client, otherwise false */ static inline bool is_privileged(pcmk__client_t *c, const char *op) { if ((c == NULL) || pcmk_is_set(c->flags, pcmk__client_privileged)) { return true; } else { crm_warn("Rejecting IPC request '%s' from unprivileged client %s", crm_str(op), pcmk__client_name(c)); return false; } } static void handle_request(pcmk__client_t *client, uint32_t id, uint32_t flags, xmlNode *request, const char *remote_peer) { int call_options = 0; int rc = -EOPNOTSUPP; xmlNode *data = NULL; bool need_reply = true; pcmk__action_result_t result = PCMK__UNKNOWN_RESULT; const char *op = crm_element_value(request, F_STONITH_OPERATION); const char *client_id = crm_element_value(request, F_STONITH_CLIENTID); crm_element_value_int(request, F_STONITH_CALLOPTS, &call_options); if (pcmk_is_set(call_options, st_opt_sync_call) && (client != NULL)) { CRM_ASSERT(client->request_id == id); } if (pcmk__str_eq(op, CRM_OP_REGISTER, pcmk__str_none)) { xmlNode *reply = create_xml_node(NULL, "reply"); CRM_ASSERT(client); crm_xml_add(reply, F_STONITH_OPERATION, CRM_OP_REGISTER); crm_xml_add(reply, F_STONITH_CLIENTID, client->id); pcmk__ipc_send_xml(client, id, reply, flags); client->request_id = 0; free_xml(reply); rc = pcmk_ok; need_reply = false; } else if (pcmk__str_eq(op, STONITH_OP_EXEC, pcmk__str_none)) { execute_agent_action(request, &result); need_reply = (result.execution_status != PCMK_EXEC_PENDING); rc = pcmk_rc2legacy(stonith__result2rc(&result)); } else if (pcmk__str_eq(op, STONITH_OP_TIMEOUT_UPDATE, pcmk__str_none)) { const char *call_id = crm_element_value(request, F_STONITH_CALLID); const char *client_id = crm_element_value(request, F_STONITH_CLIENTID); int op_timeout = 0; crm_element_value_int(request, F_STONITH_TIMEOUT, &op_timeout); do_stonith_async_timeout_update(client_id, call_id, op_timeout); rc = pcmk_ok; need_reply = false; } else if (pcmk__str_eq(op, STONITH_OP_QUERY, pcmk__str_none)) { if (remote_peer) { create_remote_stonith_op(client_id, request, TRUE); /* Record it for the future notification */ } /* Delete the DC node RELAY operation. */ remove_relay_op(request); stonith_query(request, remote_peer, client_id, call_options); rc = pcmk_ok; need_reply = false; } else if (pcmk__str_eq(op, T_STONITH_NOTIFY, pcmk__str_none)) { const char *flag_name = NULL; CRM_ASSERT(client); flag_name = crm_element_value(request, F_STONITH_NOTIFY_ACTIVATE); if (flag_name) { crm_debug("Enabling %s callbacks for client %s", flag_name, pcmk__client_name(client)); pcmk__set_client_flags(client, get_stonith_flag(flag_name)); } flag_name = crm_element_value(request, F_STONITH_NOTIFY_DEACTIVATE); if (flag_name) { crm_debug("Disabling %s callbacks for client %s", flag_name, pcmk__client_name(client)); pcmk__clear_client_flags(client, get_stonith_flag(flag_name)); } pcmk__ipc_send_ack(client, id, flags, "ack", CRM_EX_OK); rc = pcmk_ok; need_reply = false; } else if (pcmk__str_eq(op, STONITH_OP_RELAY, pcmk__str_none)) { xmlNode *dev = get_xpath_object("//@" F_STONITH_TARGET, request, LOG_TRACE); crm_notice("Received forwarded fencing request from " "%s %s to fence (%s) peer %s", ((client == NULL)? "peer" : "client"), ((client == NULL)? remote_peer : pcmk__client_name(client)), crm_element_value(dev, F_STONITH_ACTION), crm_element_value(dev, F_STONITH_TARGET)); if (initiate_remote_stonith_op(NULL, request, FALSE) == NULL) { rc = -EPROTO; } else { rc = -EINPROGRESS; need_reply = false; } } else if (pcmk__str_eq(op, STONITH_OP_FENCE, pcmk__str_none)) { if ((remote_peer != NULL) || stand_alone) { fence_locally(request, &result); rc = pcmk_rc2legacy(stonith__result2rc(&result)); } else if (pcmk_is_set(call_options, st_opt_manual_ack)) { switch (fenced_handle_manual_confirmation(client, request)) { case pcmk_rc_ok: rc = pcmk_ok; break; case EINPROGRESS: rc = -EINPROGRESS; break; default: rc = -EPROTO; break; } } else { const char *alternate_host = NULL; xmlNode *dev = get_xpath_object("//@" F_STONITH_TARGET, request, LOG_TRACE); const char *target = crm_element_value(dev, F_STONITH_TARGET); const char *action = crm_element_value(dev, F_STONITH_ACTION); const char *device = crm_element_value(dev, F_STONITH_DEVICE); if (client) { int tolerance = 0; crm_notice("Client %s wants to fence (%s) %s using %s", pcmk__client_name(client), action, target, (device? device : "any device")); crm_element_value_int(dev, F_STONITH_TOLERANCE, &tolerance); if (stonith_check_fence_tolerance(tolerance, target, action)) { rc = pcmk_ok; goto done; } } else { crm_notice("Peer %s wants to fence (%s) '%s' with device '%s'", remote_peer, action, target, device ? device : "(any)"); } alternate_host = check_alternate_host(target); if (alternate_host && client) { const char *client_id = NULL; remote_fencing_op_t *op = NULL; crm_notice("Forwarding self-fencing request to peer %s " "due to topology", alternate_host); if (client->id) { client_id = client->id; } else { client_id = crm_element_value(request, F_STONITH_CLIENTID); } /* Create an operation for RELAY and send the ID in the RELAY message. */ /* When a QUERY response is received, delete the RELAY operation to avoid the existence of duplicate operations. */ op = create_remote_stonith_op(client_id, request, FALSE); crm_xml_add(request, F_STONITH_OPERATION, STONITH_OP_RELAY); crm_xml_add(request, F_STONITH_CLIENTID, client->id); crm_xml_add(request, F_STONITH_REMOTE_OP_ID, op->id); send_cluster_message(crm_get_peer(0, alternate_host), crm_msg_stonith_ng, request, FALSE); rc = -EINPROGRESS; } else if (initiate_remote_stonith_op(client, request, FALSE) == NULL) { rc = -EPROTO; } else { rc = -EINPROGRESS; } } need_reply = (rc != -EINPROGRESS); } else if (pcmk__str_eq(op, STONITH_OP_FENCE_HISTORY, pcmk__str_none)) { stonith_fence_history(request, &data, remote_peer, call_options); rc = pcmk_ok; if (pcmk_is_set(call_options, st_opt_discard_reply)) { /* we don't expect answers to the broadcast * we might have sent out */ rc = pcmk_ok; need_reply = false; } } else if (pcmk__str_eq(op, STONITH_OP_DEVICE_ADD, pcmk__str_none)) { const char *device_id = NULL; if (is_privileged(client, op)) { rc = stonith_device_register(request, &device_id, FALSE); } else { rc = -EACCES; } do_stonith_notify_device(op, rc, device_id); } else if (pcmk__str_eq(op, STONITH_OP_DEVICE_DEL, pcmk__str_none)) { xmlNode *dev = get_xpath_object("//" F_STONITH_DEVICE, request, LOG_ERR); const char *device_id = crm_element_value(dev, XML_ATTR_ID); if (is_privileged(client, op)) { stonith_device_remove(device_id, false); rc = pcmk_ok; } else { rc = -EACCES; } do_stonith_notify_device(op, rc, device_id); } else if (pcmk__str_eq(op, STONITH_OP_LEVEL_ADD, pcmk__str_none)) { char *device_id = NULL; if (is_privileged(client, op)) { - rc = stonith_level_register(request, &device_id); + fenced_register_level(request, &device_id, &result); + rc = pcmk_rc2legacy(stonith__result2rc(&result)); } else { rc = -EACCES; } do_stonith_notify_level(op, rc, device_id); free(device_id); } else if (pcmk__str_eq(op, STONITH_OP_LEVEL_DEL, pcmk__str_none)) { char *device_id = NULL; if (is_privileged(client, op)) { rc = stonith_level_remove(request, &device_id); } else { rc = -EACCES; } do_stonith_notify_level(op, rc, device_id); } else if(pcmk__str_eq(op, CRM_OP_RM_NODE_CACHE, pcmk__str_casei)) { int node_id = 0; const char *name = NULL; crm_element_value_int(request, XML_ATTR_ID, &node_id); name = crm_element_value(request, XML_ATTR_UNAME); reap_crm_member(node_id, name); rc = pcmk_ok; need_reply = false; } else { crm_err("Unknown IPC request %s from %s %s", op, ((client == NULL)? "peer" : "client"), ((client == NULL)? remote_peer : pcmk__client_name(client))); } done: // Reply if result is known if (need_reply) { xmlNode *reply = stonith_construct_reply(request, result.action_stdout, data, rc); stonith_send_reply(reply, call_options, remote_peer, client_id); free_xml(reply); } free_xml(data); crm_debug("Processed %s request from %s %s: %s (rc=%d)", op, ((client == NULL)? "peer" : "client"), ((client == NULL)? remote_peer : pcmk__client_name(client)), ((rc > 0)? "" : pcmk_strerror(rc)), rc); pcmk__reset_result(&result); } static void handle_reply(pcmk__client_t *client, xmlNode *request, const char *remote_peer) { // Copy, because request might be freed before we want to log this char *op = crm_element_value_copy(request, F_STONITH_OPERATION); if (pcmk__str_eq(op, STONITH_OP_QUERY, pcmk__str_none)) { process_remote_stonith_query(request); } else if (pcmk__str_any_of(op, T_STONITH_NOTIFY, STONITH_OP_FENCE, NULL)) { fenced_process_fencing_reply(request); } else { crm_err("Ignoring unknown %s reply from %s %s", crm_str(op), ((client == NULL)? "peer" : "client"), ((client == NULL)? remote_peer : pcmk__client_name(client))); crm_log_xml_warn(request, "UnknownOp"); free(op); return; } crm_debug("Processed %s reply from %s %s", op, ((client == NULL)? "peer" : "client"), ((client == NULL)? remote_peer : pcmk__client_name(client))); free(op); } /*! * \internal * \brief Handle a message from an IPC client or CPG peer * * \param[in] client If not NULL, IPC client that sent message * \param[in] id If from IPC client, IPC message ID * \param[in] flags Message flags * \param[in] message Message XML * \param[in] remote_peer If not NULL, CPG peer that sent message */ void stonith_command(pcmk__client_t *client, uint32_t id, uint32_t flags, xmlNode *message, const char *remote_peer) { int call_options = st_opt_none; bool is_reply = get_xpath_object("//" T_STONITH_REPLY, message, LOG_NEVER) != NULL; crm_element_value_int(message, F_STONITH_CALLOPTS, &call_options); crm_debug("Processing %ssynchronous %s %s %u from %s %s", pcmk_is_set(call_options, st_opt_sync_call)? "" : "a", crm_element_value(message, F_STONITH_OPERATION), (is_reply? "reply" : "request"), id, ((client == NULL)? "peer" : "client"), ((client == NULL)? remote_peer : pcmk__client_name(client))); if (pcmk_is_set(call_options, st_opt_sync_call)) { CRM_ASSERT(client == NULL || client->request_id == id); } if (is_reply) { handle_reply(client, message, remote_peer); } else { handle_request(client, id, flags, message, remote_peer); } } diff --git a/daemons/fenced/pacemaker-fenced.c b/daemons/fenced/pacemaker-fenced.c index 0a8b3bf6f2..469304f67c 100644 --- a/daemons/fenced/pacemaker-fenced.c +++ b/daemons/fenced/pacemaker-fenced.c @@ -1,1668 +1,1669 @@ /* * Copyright 2009-2021 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include #include #include // PRIu32, PRIx32 #include #include #include #include #include #include #include #include #include #include #include #include #include #include char *stonith_our_uname = NULL; long stonith_watchdog_timeout_ms = 0; GList *stonith_watchdog_targets = NULL; static GMainLoop *mainloop = NULL; gboolean stand_alone = FALSE; static gboolean no_cib_connect = FALSE; static gboolean stonith_shutdown_flag = FALSE; static qb_ipcs_service_t *ipcs = NULL; static xmlNode *local_cib = NULL; static pe_working_set_t *fenced_data_set = NULL; static cib_t *cib_api = NULL; static pcmk__output_t *out = NULL; pcmk__supported_format_t formats[] = { PCMK__SUPPORTED_FORMAT_LOG, PCMK__SUPPORTED_FORMAT_NONE, PCMK__SUPPORTED_FORMAT_TEXT, { NULL, NULL, NULL } }; static void stonith_shutdown(int nsig); static void stonith_cleanup(void); static int32_t st_ipc_accept(qb_ipcs_connection_t * c, uid_t uid, gid_t gid) { if (stonith_shutdown_flag) { crm_info("Ignoring new client [%d] during shutdown", pcmk__client_pid(c)); return -EPERM; } if (pcmk__new_client(c, uid, gid) == NULL) { return -EIO; } return 0; } /* Exit code means? */ static int32_t st_ipc_dispatch(qb_ipcs_connection_t * qbc, void *data, size_t size) { uint32_t id = 0; uint32_t flags = 0; int call_options = 0; xmlNode *request = NULL; pcmk__client_t *c = pcmk__find_client(qbc); const char *op = NULL; if (c == NULL) { crm_info("Invalid client: %p", qbc); return 0; } request = pcmk__client_data2xml(c, data, &id, &flags); if (request == NULL) { pcmk__ipc_send_ack(c, id, flags, "nack", CRM_EX_PROTOCOL); return 0; } op = crm_element_value(request, F_CRM_TASK); if(pcmk__str_eq(op, CRM_OP_RM_NODE_CACHE, pcmk__str_casei)) { crm_xml_add(request, F_TYPE, T_STONITH_NG); crm_xml_add(request, F_STONITH_OPERATION, op); crm_xml_add(request, F_STONITH_CLIENTID, c->id); crm_xml_add(request, F_STONITH_CLIENTNAME, pcmk__client_name(c)); crm_xml_add(request, F_STONITH_CLIENTNODE, stonith_our_uname); send_cluster_message(NULL, crm_msg_stonith_ng, request, FALSE); free_xml(request); return 0; } if (c->name == NULL) { const char *value = crm_element_value(request, F_STONITH_CLIENTNAME); if (value == NULL) { value = "unknown"; } c->name = crm_strdup_printf("%s.%u", value, c->pid); } crm_element_value_int(request, F_STONITH_CALLOPTS, &call_options); crm_trace("Flags 0x%08" PRIx32 "/0x%08x for command %" PRIu32 " from client %s", flags, call_options, id, pcmk__client_name(c)); if (pcmk_is_set(call_options, st_opt_sync_call)) { CRM_ASSERT(flags & crm_ipc_client_response); CRM_LOG_ASSERT(c->request_id == 0); /* This means the client has two synchronous events in-flight */ c->request_id = id; /* Reply only to the last one */ } crm_xml_add(request, F_STONITH_CLIENTID, c->id); crm_xml_add(request, F_STONITH_CLIENTNAME, pcmk__client_name(c)); crm_xml_add(request, F_STONITH_CLIENTNODE, stonith_our_uname); stonith_command(c, id, flags, request, NULL); free_xml(request); return 0; } /* Error code means? */ static int32_t st_ipc_closed(qb_ipcs_connection_t * c) { pcmk__client_t *client = pcmk__find_client(c); if (client == NULL) { return 0; } crm_trace("Connection %p closed", c); pcmk__free_client(client); /* 0 means: yes, go ahead and destroy the connection */ return 0; } static void st_ipc_destroy(qb_ipcs_connection_t * c) { crm_trace("Connection %p destroyed", c); st_ipc_closed(c); } static void stonith_peer_callback(xmlNode * msg, void *private_data) { const char *remote_peer = crm_element_value(msg, F_ORIG); const char *op = crm_element_value(msg, F_STONITH_OPERATION); if (pcmk__str_eq(op, "poke", pcmk__str_none)) { return; } crm_log_xml_trace(msg, "Peer[inbound]"); stonith_command(NULL, 0, 0, msg, remote_peer); } #if SUPPORT_COROSYNC static void stonith_peer_ais_callback(cpg_handle_t handle, const struct cpg_name *groupName, uint32_t nodeid, uint32_t pid, void *msg, size_t msg_len) { uint32_t kind = 0; xmlNode *xml = NULL; const char *from = NULL; char *data = pcmk_message_common_cs(handle, nodeid, pid, msg, &kind, &from); if(data == NULL) { return; } if (kind == crm_class_cluster) { xml = string2xml(data); if (xml == NULL) { crm_err("Invalid XML: '%.120s'", data); free(data); return; } crm_xml_add(xml, F_ORIG, from); /* crm_xml_add_int(xml, F_SEQ, wrapper->id); */ stonith_peer_callback(xml, NULL); } free_xml(xml); free(data); return; } static void stonith_peer_cs_destroy(gpointer user_data) { crm_crit("Lost connection to cluster layer, shutting down"); stonith_shutdown(0); } #endif void do_local_reply(xmlNode * notify_src, const char *client_id, gboolean sync_reply, gboolean from_peer) { /* send callback to originating child */ pcmk__client_t *client_obj = NULL; int local_rc = pcmk_rc_ok; crm_trace("Sending response"); client_obj = pcmk__find_client_by_id(client_id); crm_trace("Sending callback to request originator"); if (client_obj == NULL) { local_rc = EPROTO; crm_trace("No client to sent the response to. F_STONITH_CLIENTID not set."); } else { int rid = 0; if (sync_reply) { CRM_LOG_ASSERT(client_obj->request_id); rid = client_obj->request_id; client_obj->request_id = 0; crm_trace("Sending response %d to client %s%s", rid, pcmk__client_name(client_obj), (from_peer? " (originator of delegated request)" : "")); } else { crm_trace("Sending an event to client %s%s", pcmk__client_name(client_obj), (from_peer? " (originator of delegated request)" : "")); } local_rc = pcmk__ipc_send_xml(client_obj, rid, notify_src, (sync_reply? crm_ipc_flags_none : crm_ipc_server_event)); } if ((local_rc != pcmk_rc_ok) && (client_obj != NULL)) { crm_warn("%s reply to client %s failed: %s", (sync_reply? "Synchronous" : "Asynchronous"), pcmk__client_name(client_obj), pcmk_rc_str(local_rc)); } } uint64_t get_stonith_flag(const char *name) { if (pcmk__str_eq(name, T_STONITH_NOTIFY_FENCE, pcmk__str_casei)) { return st_callback_notify_fence; } else if (pcmk__str_eq(name, STONITH_OP_DEVICE_ADD, pcmk__str_casei)) { return st_callback_device_add; } else if (pcmk__str_eq(name, STONITH_OP_DEVICE_DEL, pcmk__str_casei)) { return st_callback_device_del; } else if (pcmk__str_eq(name, T_STONITH_NOTIFY_HISTORY, pcmk__str_casei)) { return st_callback_notify_history; } else if (pcmk__str_eq(name, T_STONITH_NOTIFY_HISTORY_SYNCED, pcmk__str_casei)) { return st_callback_notify_history_synced; } return st_callback_unknown; } static void stonith_notify_client(gpointer key, gpointer value, gpointer user_data) { xmlNode *update_msg = user_data; pcmk__client_t *client = value; const char *type = NULL; CRM_CHECK(client != NULL, return); CRM_CHECK(update_msg != NULL, return); type = crm_element_value(update_msg, F_SUBTYPE); CRM_CHECK(type != NULL, crm_log_xml_err(update_msg, "notify"); return); if (client->ipcs == NULL) { crm_trace("Skipping client with NULL channel"); return; } if (pcmk_is_set(client->flags, get_stonith_flag(type))) { int rc = pcmk__ipc_send_xml(client, 0, update_msg, crm_ipc_server_event|crm_ipc_server_error); if (rc != pcmk_rc_ok) { crm_warn("%s notification of client %s failed: %s " CRM_XS " id=%.8s rc=%d", type, pcmk__client_name(client), pcmk_rc_str(rc), client->id, rc); } else { crm_trace("Sent %s notification to client %s", type, pcmk__client_name(client)); } } } void do_stonith_async_timeout_update(const char *client_id, const char *call_id, int timeout) { pcmk__client_t *client = NULL; xmlNode *notify_data = NULL; if (!timeout || !call_id || !client_id) { return; } client = pcmk__find_client_by_id(client_id); if (!client) { return; } notify_data = create_xml_node(NULL, T_STONITH_TIMEOUT_VALUE); crm_xml_add(notify_data, F_TYPE, T_STONITH_TIMEOUT_VALUE); crm_xml_add(notify_data, F_STONITH_CALLID, call_id); crm_xml_add_int(notify_data, F_STONITH_TIMEOUT, timeout); crm_trace("timeout update is %d for client %s and call id %s", timeout, client_id, call_id); if (client) { pcmk__ipc_send_xml(client, 0, notify_data, crm_ipc_server_event); } free_xml(notify_data); } void do_stonith_notify(const char *type, int result, xmlNode *data) { /* TODO: Standardize the contents of data */ xmlNode *update_msg = create_xml_node(NULL, "notify"); CRM_CHECK(type != NULL,;); crm_xml_add(update_msg, F_TYPE, T_STONITH_NOTIFY); crm_xml_add(update_msg, F_SUBTYPE, type); crm_xml_add(update_msg, F_STONITH_OPERATION, type); crm_xml_add_int(update_msg, F_STONITH_RC, result); if (data != NULL) { add_message_xml(update_msg, F_STONITH_CALLDATA, data); } crm_trace("Notifying clients"); pcmk__foreach_ipc_client(stonith_notify_client, update_msg); free_xml(update_msg); crm_trace("Notify complete"); } static void do_stonith_notify_config(const char *op, int rc, const char *desc, int active) { xmlNode *notify_data = create_xml_node(NULL, op); CRM_CHECK(notify_data != NULL, return); crm_xml_add(notify_data, F_STONITH_DEVICE, desc); crm_xml_add_int(notify_data, F_STONITH_ACTIVE, active); do_stonith_notify(op, rc, notify_data); free_xml(notify_data); } void do_stonith_notify_device(const char *op, int rc, const char *desc) { do_stonith_notify_config(op, rc, desc, g_hash_table_size(device_list)); } void do_stonith_notify_level(const char *op, int rc, const char *desc) { do_stonith_notify_config(op, rc, desc, g_hash_table_size(topology)); } static void topology_remove_helper(const char *node, int level) { int rc; char *desc = NULL; xmlNode *data = create_xml_node(NULL, XML_TAG_FENCING_LEVEL); crm_xml_add(data, F_STONITH_ORIGIN, __func__); crm_xml_add_int(data, XML_ATTR_STONITH_INDEX, level); crm_xml_add(data, XML_ATTR_STONITH_TARGET, node); rc = stonith_level_remove(data, &desc); do_stonith_notify_level(STONITH_OP_LEVEL_DEL, rc, desc); free_xml(data); free(desc); } static void remove_cib_device(xmlXPathObjectPtr xpathObj) { int max = numXpathResults(xpathObj), lpc = 0; for (lpc = 0; lpc < max; lpc++) { const char *rsc_id = NULL; const char *standard = NULL; xmlNode *match = getXpathResult(xpathObj, lpc); CRM_LOG_ASSERT(match != NULL); if(match != NULL) { standard = crm_element_value(match, XML_AGENT_ATTR_CLASS); } if (!pcmk__str_eq(standard, PCMK_RESOURCE_CLASS_STONITH, pcmk__str_casei)) { continue; } rsc_id = crm_element_value(match, XML_ATTR_ID); stonith_device_remove(rsc_id, true); } } static void handle_topology_change(xmlNode *match, bool remove) { - int rc; char *desc = NULL; + pcmk__action_result_t result = PCMK__UNKNOWN_RESULT; CRM_CHECK(match != NULL, return); crm_trace("Updating %s", ID(match)); if(remove) { int index = 0; char *key = stonith_level_key(match, -1); crm_element_value_int(match, XML_ATTR_STONITH_INDEX, &index); topology_remove_helper(key, index); free(key); } - rc = stonith_level_register(match, &desc); - do_stonith_notify_level(STONITH_OP_LEVEL_ADD, rc, desc); - + fenced_register_level(match, &desc, &result); + do_stonith_notify_level(STONITH_OP_LEVEL_ADD, + pcmk_rc2legacy(stonith__result2rc(&result)), desc); + pcmk__reset_result(&result); free(desc); } static void remove_fencing_topology(xmlXPathObjectPtr xpathObj) { int max = numXpathResults(xpathObj), lpc = 0; for (lpc = 0; lpc < max; lpc++) { xmlNode *match = getXpathResult(xpathObj, lpc); CRM_LOG_ASSERT(match != NULL); if (match && crm_element_value(match, XML_DIFF_MARKER)) { /* Deletion */ int index = 0; char *target = stonith_level_key(match, -1); crm_element_value_int(match, XML_ATTR_STONITH_INDEX, &index); if (target == NULL) { crm_err("Invalid fencing target in element %s", ID(match)); } else if (index <= 0) { crm_err("Invalid level for %s in element %s", target, ID(match)); } else { topology_remove_helper(target, index); } /* } else { Deal with modifications during the 'addition' stage */ } } } static void register_fencing_topology(xmlXPathObjectPtr xpathObj) { int max = numXpathResults(xpathObj), lpc = 0; for (lpc = 0; lpc < max; lpc++) { xmlNode *match = getXpathResult(xpathObj, lpc); handle_topology_change(match, TRUE); } } /* Fencing */ static void fencing_topology_init(void) { xmlXPathObjectPtr xpathObj = NULL; const char *xpath = "//" XML_TAG_FENCING_LEVEL; crm_trace("Full topology refresh"); free_topology_list(); init_topology_list(); /* Grab everything */ xpathObj = xpath_search(local_cib, xpath); register_fencing_topology(xpathObj); freeXpathObject(xpathObj); } #define rsc_name(x) x->clone_name?x->clone_name:x->id /*! * \internal * \brief Check whether our uname is in a resource's allowed node list * * \param[in] rsc Resource to check * * \return Pointer to node object if found, NULL otherwise */ static pe_node_t * our_node_allowed_for(pe_resource_t *rsc) { GHashTableIter iter; pe_node_t *node = NULL; if (rsc && stonith_our_uname) { g_hash_table_iter_init(&iter, rsc->allowed_nodes); while (g_hash_table_iter_next(&iter, NULL, (void **)&node)) { if (node && strcmp(node->details->uname, stonith_our_uname) == 0) { break; } node = NULL; } } return node; } static void watchdog_device_update(void) { if (stonith_watchdog_timeout_ms > 0) { if (!g_hash_table_lookup(device_list, STONITH_WATCHDOG_ID) && !stonith_watchdog_targets) { /* getting here watchdog-fencing enabled, no device there yet and reason isn't stonith_watchdog_targets preventing that */ int rc; xmlNode *xml; xml = create_device_registration_xml( STONITH_WATCHDOG_ID, st_namespace_internal, STONITH_WATCHDOG_AGENT, NULL, /* stonith_device_register will add our own name as PCMK_STONITH_HOST_LIST param so we can skip that here */ NULL); rc = stonith_device_register(xml, NULL, TRUE); free_xml(xml); if (rc != pcmk_ok) { crm_crit("Cannot register watchdog pseudo fence agent"); crm_exit(CRM_EX_FATAL); } } } else { /* be silent if no device - todo parameter to stonith_device_remove */ if (g_hash_table_lookup(device_list, STONITH_WATCHDOG_ID)) { stonith_device_remove(STONITH_WATCHDOG_ID, true); } } } static void update_stonith_watchdog_timeout_ms(xmlNode *cib) { xmlNode *stonith_enabled_xml = NULL; bool stonith_enabled = false; int rc = pcmk_rc_ok; long timeout_ms = 0; stonith_enabled_xml = get_xpath_object("//nvpair[@name='stonith-enabled']", cib, LOG_NEVER); rc = pcmk__xe_get_bool_attr(stonith_enabled_xml, XML_NVPAIR_ATTR_VALUE, &stonith_enabled); if (rc != pcmk_rc_ok || stonith_enabled) { xmlNode *stonith_watchdog_xml = NULL; const char *value = NULL; stonith_watchdog_xml = get_xpath_object("//nvpair[@name='stonith-watchdog-timeout']", cib, LOG_NEVER); if (stonith_watchdog_xml) { value = crm_element_value(stonith_watchdog_xml, XML_NVPAIR_ATTR_VALUE); } if (value) { timeout_ms = crm_get_msec(value); } if (timeout_ms < 0) { timeout_ms = pcmk__auto_watchdog_timeout(); } } stonith_watchdog_timeout_ms = timeout_ms; } /*! * \internal * \brief If a resource or any of its children are STONITH devices, update their * definitions given a cluster working set. * * \param[in] rsc Resource to check * \param[in] data_set Cluster working set with device information */ static void cib_device_update(pe_resource_t *rsc, pe_working_set_t *data_set) { pe_node_t *node = NULL; const char *value = NULL; const char *rclass = NULL; pe_node_t *parent = NULL; /* If this is a complex resource, check children rather than this resource itself. */ if(rsc->children) { GList *gIter = NULL; for (gIter = rsc->children; gIter != NULL; gIter = gIter->next) { cib_device_update(gIter->data, data_set); if(pe_rsc_is_clone(rsc)) { crm_trace("Only processing one copy of the clone %s", rsc->id); break; } } return; } /* We only care about STONITH resources. */ rclass = crm_element_value(rsc->xml, XML_AGENT_ATTR_CLASS); if (!pcmk__str_eq(rclass, PCMK_RESOURCE_CLASS_STONITH, pcmk__str_casei)) { return; } /* If this STONITH resource is disabled, remove it. */ if (pe__resource_is_disabled(rsc)) { crm_info("Device %s has been disabled", rsc->id); return; } /* if watchdog-fencing is disabled handle any watchdog-fence resource as if it was disabled */ if ((stonith_watchdog_timeout_ms <= 0) && pcmk__str_eq(rsc->id, STONITH_WATCHDOG_ID, pcmk__str_none)) { crm_info("Watchdog-fencing disabled thus handling " "device %s as disabled", rsc->id); return; } /* Check whether our node is allowed for this resource (and its parent if in a group) */ node = our_node_allowed_for(rsc); if (rsc->parent && (rsc->parent->variant == pe_group)) { parent = our_node_allowed_for(rsc->parent); } if(node == NULL) { /* Our node is disallowed, so remove the device */ GHashTableIter iter; crm_info("Device %s has been disabled on %s: unknown", rsc->id, stonith_our_uname); g_hash_table_iter_init(&iter, rsc->allowed_nodes); while (g_hash_table_iter_next(&iter, NULL, (void **)&node)) { crm_trace("Available: %s = %d", node->details->uname, node->weight); } return; } else if(node->weight < 0 || (parent && parent->weight < 0)) { /* Our node (or its group) is disallowed by score, so remove the device */ char *score = score2char((node->weight < 0) ? node->weight : parent->weight); crm_info("Device %s has been disabled on %s: score=%s", rsc->id, stonith_our_uname, score); free(score); return; } else { /* Our node is allowed, so update the device information */ int rc; xmlNode *data; GHashTable *rsc_params = NULL; GHashTableIter gIter; stonith_key_value_t *params = NULL; const char *name = NULL; const char *agent = crm_element_value(rsc->xml, XML_EXPR_ATTR_TYPE); const char *rsc_provides = NULL; crm_debug("Device %s is allowed on %s: score=%d", rsc->id, stonith_our_uname, node->weight); rsc_params = pe_rsc_params(rsc, node, data_set); get_meta_attributes(rsc->meta, rsc, node, data_set); rsc_provides = g_hash_table_lookup(rsc->meta, PCMK_STONITH_PROVIDES); g_hash_table_iter_init(&gIter, rsc_params); while (g_hash_table_iter_next(&gIter, (gpointer *) & name, (gpointer *) & value)) { if (!name || !value) { continue; } params = stonith_key_value_add(params, name, value); crm_trace(" %s=%s", name, value); } data = create_device_registration_xml(rsc_name(rsc), st_namespace_any, agent, params, rsc_provides); stonith_key_value_freeall(params, 1, 1); rc = stonith_device_register(data, NULL, TRUE); CRM_ASSERT(rc == pcmk_ok); free_xml(data); } } /*! * \internal * \brief Update all STONITH device definitions based on current CIB */ static void cib_devices_update(void) { GHashTableIter iter; stonith_device_t *device = NULL; crm_info("Updating devices to version %s.%s.%s", crm_element_value(local_cib, XML_ATTR_GENERATION_ADMIN), crm_element_value(local_cib, XML_ATTR_GENERATION), crm_element_value(local_cib, XML_ATTR_NUMUPDATES)); CRM_ASSERT(fenced_data_set != NULL); fenced_data_set->input = local_cib; fenced_data_set->now = crm_time_new(NULL); fenced_data_set->localhost = stonith_our_uname; pe__set_working_set_flags(fenced_data_set, pe_flag_quick_location); cluster_status(fenced_data_set); pcmk__schedule_actions(fenced_data_set, NULL, NULL); g_hash_table_iter_init(&iter, device_list); while (g_hash_table_iter_next(&iter, NULL, (void **)&device)) { if (device->cib_registered) { device->dirty = TRUE; } } /* have list repopulated if cib has a watchdog-fencing-resource TODO: keep a cached list for queries happening while we are refreshing */ g_list_free_full(stonith_watchdog_targets, free); stonith_watchdog_targets = NULL; g_list_foreach(fenced_data_set->resources, (GFunc) cib_device_update, fenced_data_set); g_hash_table_iter_init(&iter, device_list); while (g_hash_table_iter_next(&iter, NULL, (void **)&device)) { if (device->dirty) { g_hash_table_iter_remove(&iter); } } fenced_data_set->input = NULL; // Wasn't a copy, so don't let API free it pe_reset_working_set(fenced_data_set); } static void update_cib_stonith_devices_v2(const char *event, xmlNode * msg) { xmlNode *change = NULL; char *reason = NULL; bool needs_update = FALSE; xmlNode *patchset = get_message_xml(msg, F_CIB_UPDATE_RESULT); for (change = pcmk__xml_first_child(patchset); change != NULL; change = pcmk__xml_next(change)) { const char *op = crm_element_value(change, XML_DIFF_OP); const char *xpath = crm_element_value(change, XML_DIFF_PATH); const char *shortpath = NULL; if ((op == NULL) || (strcmp(op, "move") == 0) || strstr(xpath, "/"XML_CIB_TAG_STATUS)) { continue; } else if (pcmk__str_eq(op, "delete", pcmk__str_casei) && strstr(xpath, "/"XML_CIB_TAG_RESOURCE)) { const char *rsc_id = NULL; char *search = NULL; char *mutable = NULL; if (strstr(xpath, XML_TAG_ATTR_SETS) || strstr(xpath, XML_TAG_META_SETS)) { needs_update = TRUE; reason = strdup("(meta) attribute deleted from resource"); break; } mutable = strdup(xpath); rsc_id = strstr(mutable, "primitive[@id=\'"); if (rsc_id != NULL) { rsc_id += strlen("primitive[@id=\'"); search = strchr(rsc_id, '\''); } if (search != NULL) { *search = 0; stonith_device_remove(rsc_id, true); /* watchdog_device_update called afterwards to fall back to implicit definition if needed */ } else { crm_warn("Ignoring malformed CIB update (resource deletion)"); } free(mutable); } else if (strstr(xpath, "/"XML_CIB_TAG_RESOURCES) || strstr(xpath, "/"XML_CIB_TAG_CONSTRAINTS) || strstr(xpath, "/"XML_CIB_TAG_RSCCONFIG)) { shortpath = strrchr(xpath, '/'); CRM_ASSERT(shortpath); reason = crm_strdup_printf("%s %s", op, shortpath+1); needs_update = TRUE; break; } } if(needs_update) { crm_info("Updating device list from CIB: %s", reason); cib_devices_update(); } else { crm_trace("No updates for device list found in CIB"); } free(reason); } static void update_cib_stonith_devices_v1(const char *event, xmlNode * msg) { const char *reason = "none"; gboolean needs_update = FALSE; xmlXPathObjectPtr xpath_obj = NULL; /* process new constraints */ xpath_obj = xpath_search(msg, "//" F_CIB_UPDATE_RESULT "//" XML_CONS_TAG_RSC_LOCATION); if (numXpathResults(xpath_obj) > 0) { int max = numXpathResults(xpath_obj), lpc = 0; /* Safest and simplest to always recompute */ needs_update = TRUE; reason = "new location constraint"; for (lpc = 0; lpc < max; lpc++) { xmlNode *match = getXpathResult(xpath_obj, lpc); crm_log_xml_trace(match, "new constraint"); } } freeXpathObject(xpath_obj); /* process deletions */ xpath_obj = xpath_search(msg, "//" F_CIB_UPDATE_RESULT "//" XML_TAG_DIFF_REMOVED "//" XML_CIB_TAG_RESOURCE); if (numXpathResults(xpath_obj) > 0) { remove_cib_device(xpath_obj); } freeXpathObject(xpath_obj); /* process additions */ xpath_obj = xpath_search(msg, "//" F_CIB_UPDATE_RESULT "//" XML_TAG_DIFF_ADDED "//" XML_CIB_TAG_RESOURCE); if (numXpathResults(xpath_obj) > 0) { int max = numXpathResults(xpath_obj), lpc = 0; for (lpc = 0; lpc < max; lpc++) { const char *rsc_id = NULL; const char *standard = NULL; xmlNode *match = getXpathResult(xpath_obj, lpc); rsc_id = crm_element_value(match, XML_ATTR_ID); standard = crm_element_value(match, XML_AGENT_ATTR_CLASS); if (!pcmk__str_eq(standard, PCMK_RESOURCE_CLASS_STONITH, pcmk__str_casei)) { continue; } crm_trace("Fencing resource %s was added or modified", rsc_id); reason = "new resource"; needs_update = TRUE; } } freeXpathObject(xpath_obj); if(needs_update) { crm_info("Updating device list from CIB: %s", reason); cib_devices_update(); } } static void update_cib_stonith_devices(const char *event, xmlNode * msg) { int format = 1; xmlNode *patchset = get_message_xml(msg, F_CIB_UPDATE_RESULT); CRM_ASSERT(patchset); crm_element_value_int(patchset, "format", &format); switch(format) { case 1: update_cib_stonith_devices_v1(event, msg); break; case 2: update_cib_stonith_devices_v2(event, msg); break; default: crm_warn("Unknown patch format: %d", format); } } /* Needs to hold node name + attribute name + attribute value + 75 */ #define XPATH_MAX 512 /*! * \internal * \brief Check whether a node has a specific attribute name/value * * \param[in] node Name of node to check * \param[in] name Name of an attribute to look for * \param[in] value The value the named attribute needs to be set to in order to be considered a match * * \return TRUE if the locally cached CIB has the specified node attribute */ gboolean node_has_attr(const char *node, const char *name, const char *value) { char xpath[XPATH_MAX]; xmlNode *match; int n; CRM_CHECK(local_cib != NULL, return FALSE); /* Search for the node's attributes in the CIB. While the schema allows * multiple sets of instance attributes, and allows instance attributes to * use id-ref to reference values elsewhere, that is intended for resources, * so we ignore that here. */ n = snprintf(xpath, XPATH_MAX, "//" XML_CIB_TAG_NODES "/" XML_CIB_TAG_NODE "[@uname='%s']/" XML_TAG_ATTR_SETS "/" XML_CIB_TAG_NVPAIR "[@name='%s' and @value='%s']", node, name, value); match = get_xpath_object(xpath, local_cib, LOG_NEVER); CRM_CHECK(n < XPATH_MAX, return FALSE); return (match != NULL); } /*! * \internal * \brief Check whether a node does watchdog-fencing * * \param[in] node Name of node to check * * \return TRUE if node found in stonith_watchdog_targets * or stonith_watchdog_targets is empty indicating * all nodes are doing watchdog-fencing */ gboolean node_does_watchdog_fencing(const char *node) { return ((stonith_watchdog_targets == NULL) || pcmk__str_in_list(node, stonith_watchdog_targets, pcmk__str_casei)); } static void update_fencing_topology(const char *event, xmlNode * msg) { int format = 1; const char *xpath; xmlXPathObjectPtr xpathObj = NULL; xmlNode *patchset = get_message_xml(msg, F_CIB_UPDATE_RESULT); CRM_ASSERT(patchset); crm_element_value_int(patchset, "format", &format); if(format == 1) { /* Process deletions (only) */ xpath = "//" F_CIB_UPDATE_RESULT "//" XML_TAG_DIFF_REMOVED "//" XML_TAG_FENCING_LEVEL; xpathObj = xpath_search(msg, xpath); remove_fencing_topology(xpathObj); freeXpathObject(xpathObj); /* Process additions and changes */ xpath = "//" F_CIB_UPDATE_RESULT "//" XML_TAG_DIFF_ADDED "//" XML_TAG_FENCING_LEVEL; xpathObj = xpath_search(msg, xpath); register_fencing_topology(xpathObj); freeXpathObject(xpathObj); } else if(format == 2) { xmlNode *change = NULL; int add[] = { 0, 0, 0 }; int del[] = { 0, 0, 0 }; xml_patch_versions(patchset, add, del); for (change = pcmk__xml_first_child(patchset); change != NULL; change = pcmk__xml_next(change)) { const char *op = crm_element_value(change, XML_DIFF_OP); const char *xpath = crm_element_value(change, XML_DIFF_PATH); if(op == NULL) { continue; } else if(strstr(xpath, "/" XML_TAG_FENCING_LEVEL) != NULL) { /* Change to a specific entry */ crm_trace("Handling %s operation %d.%d.%d for %s", op, add[0], add[1], add[2], xpath); if(strcmp(op, "move") == 0) { continue; } else if(strcmp(op, "create") == 0) { handle_topology_change(change->children, FALSE); } else if(strcmp(op, "modify") == 0) { xmlNode *match = first_named_child(change, XML_DIFF_RESULT); if(match) { handle_topology_change(match->children, TRUE); } } else if(strcmp(op, "delete") == 0) { /* Nuclear option, all we have is the path and an id... not enough to remove a specific entry */ crm_info("Re-initializing fencing topology after %s operation %d.%d.%d for %s", op, add[0], add[1], add[2], xpath); fencing_topology_init(); return; } } else if (strstr(xpath, "/" XML_TAG_FENCING_TOPOLOGY) != NULL) { /* Change to the topology in general */ crm_info("Re-initializing fencing topology after top-level %s operation %d.%d.%d for %s", op, add[0], add[1], add[2], xpath); fencing_topology_init(); return; } else if (strstr(xpath, "/" XML_CIB_TAG_CONFIGURATION)) { /* Changes to the whole config section, possibly including the topology as a whild */ if(first_named_child(change, XML_TAG_FENCING_TOPOLOGY) == NULL) { crm_trace("Nothing for us in %s operation %d.%d.%d for %s.", op, add[0], add[1], add[2], xpath); } else if(strcmp(op, "delete") == 0 || strcmp(op, "create") == 0) { crm_info("Re-initializing fencing topology after top-level %s operation %d.%d.%d for %s.", op, add[0], add[1], add[2], xpath); fencing_topology_init(); return; } } else { crm_trace("Nothing for us in %s operation %d.%d.%d for %s", op, add[0], add[1], add[2], xpath); } } } else { crm_warn("Unknown patch format: %d", format); } } static bool have_cib_devices = FALSE; static void update_cib_cache_cb(const char *event, xmlNode * msg) { int rc = pcmk_ok; xmlNode *stonith_enabled_xml = NULL; static gboolean stonith_enabled_saved = TRUE; long timeout_ms_saved = stonith_watchdog_timeout_ms; gboolean need_full_refresh = FALSE; bool value = false; if(!have_cib_devices) { crm_trace("Skipping updates until we get a full dump"); return; } else if(msg == NULL) { crm_trace("Missing %s update", event); return; } /* Maintain a local copy of the CIB so that we have full access * to device definitions, location constraints, and node attributes */ if (local_cib != NULL) { int rc = pcmk_ok; xmlNode *patchset = NULL; crm_element_value_int(msg, F_CIB_RC, &rc); if (rc != pcmk_ok) { return; } patchset = get_message_xml(msg, F_CIB_UPDATE_RESULT); xml_log_patchset(LOG_TRACE, "Config update", patchset); rc = xml_apply_patchset(local_cib, patchset, TRUE); switch (rc) { case pcmk_ok: case -pcmk_err_old_data: break; case -pcmk_err_diff_resync: case -pcmk_err_diff_failed: crm_notice("[%s] Patch aborted: %s (%d)", event, pcmk_strerror(rc), rc); free_xml(local_cib); local_cib = NULL; break; default: crm_warn("[%s] ABORTED: %s (%d)", event, pcmk_strerror(rc), rc); free_xml(local_cib); local_cib = NULL; } } if (local_cib == NULL) { crm_trace("Re-requesting full CIB"); rc = cib_api->cmds->query(cib_api, NULL, &local_cib, cib_scope_local | cib_sync_call); if(rc != pcmk_ok) { crm_err("Couldn't retrieve the CIB: %s (%d)", pcmk_strerror(rc), rc); return; } CRM_ASSERT(local_cib != NULL); stonith_enabled_saved = FALSE; /* Trigger a full refresh below */ } pcmk__refresh_node_caches_from_cib(local_cib); update_stonith_watchdog_timeout_ms(local_cib); stonith_enabled_xml = get_xpath_object("//nvpair[@name='stonith-enabled']", local_cib, LOG_NEVER); if (pcmk__xe_get_bool_attr(stonith_enabled_xml, XML_NVPAIR_ATTR_VALUE, &value) == pcmk_rc_ok && !value) { crm_trace("Ignoring CIB updates while fencing is disabled"); stonith_enabled_saved = FALSE; } else if (stonith_enabled_saved == FALSE) { crm_info("Updating fencing device and topology lists " "now that fencing is enabled"); stonith_enabled_saved = TRUE; need_full_refresh = TRUE; } else { if (timeout_ms_saved != stonith_watchdog_timeout_ms) { need_full_refresh = TRUE; } else { update_fencing_topology(event, msg); update_cib_stonith_devices(event, msg); watchdog_device_update(); } } if (need_full_refresh) { fencing_topology_init(); cib_devices_update(); watchdog_device_update(); } } static void init_cib_cache_cb(xmlNode * msg, int call_id, int rc, xmlNode * output, void *user_data) { crm_info("Updating device list from CIB"); have_cib_devices = TRUE; local_cib = copy_xml(output); pcmk__refresh_node_caches_from_cib(local_cib); update_stonith_watchdog_timeout_ms(local_cib); fencing_topology_init(); cib_devices_update(); watchdog_device_update(); } static void stonith_shutdown(int nsig) { crm_info("Terminating with %d clients", pcmk__ipc_client_count()); stonith_shutdown_flag = TRUE; if (mainloop != NULL && g_main_loop_is_running(mainloop)) { g_main_loop_quit(mainloop); } else { stonith_cleanup(); crm_exit(CRM_EX_OK); } } static void cib_connection_destroy(gpointer user_data) { if (stonith_shutdown_flag) { crm_info("Connection to the CIB manager closed"); return; } else { crm_crit("Lost connection to the CIB manager, shutting down"); } if (cib_api) { cib_api->cmds->signoff(cib_api); } stonith_shutdown(0); } static void stonith_cleanup(void) { if (cib_api) { cib_api->cmds->del_notify_callback(cib_api, T_CIB_DIFF_NOTIFY, update_cib_cache_cb); cib_api->cmds->signoff(cib_api); } if (ipcs) { qb_ipcs_destroy(ipcs); } crm_peer_destroy(); pcmk__client_cleanup(); free_stonith_remote_op_list(); free_topology_list(); free_device_list(); free_metadata_cache(); free(stonith_our_uname); stonith_our_uname = NULL; free_xml(local_cib); local_cib = NULL; } static pcmk__cli_option_t long_options[] = { // long option, argument type, storage, short option, description, flags { "stand-alone", no_argument, 0, 's', NULL, pcmk__option_default }, { "stand-alone-w-cpg", no_argument, 0, 'c', NULL, pcmk__option_default }, { "logfile", required_argument, 0, 'l', NULL, pcmk__option_default }, { "verbose", no_argument, 0, 'V', NULL, pcmk__option_default }, { "version", no_argument, 0, '$', NULL, pcmk__option_default }, { "help", no_argument, 0, '?', NULL, pcmk__option_default }, { 0, 0, 0, 0 } }; static void setup_cib(void) { int rc, retries = 0; cib_api = cib_new(); if (cib_api == NULL) { crm_err("No connection to the CIB manager"); return; } do { sleep(retries); rc = cib_api->cmds->signon(cib_api, CRM_SYSTEM_STONITHD, cib_command); } while (rc == -ENOTCONN && ++retries < 5); if (rc != pcmk_ok) { crm_err("Could not connect to the CIB manager: %s (%d)", pcmk_strerror(rc), rc); } else if (pcmk_ok != cib_api->cmds->add_notify_callback(cib_api, T_CIB_DIFF_NOTIFY, update_cib_cache_cb)) { crm_err("Could not set CIB notification callback"); } else { rc = cib_api->cmds->query(cib_api, NULL, NULL, cib_scope_local); cib_api->cmds->register_callback(cib_api, rc, 120, FALSE, NULL, "init_cib_cache_cb", init_cib_cache_cb); cib_api->cmds->set_connection_dnotify(cib_api, cib_connection_destroy); crm_info("Watching for fencing topology changes"); } } struct qb_ipcs_service_handlers ipc_callbacks = { .connection_accept = st_ipc_accept, .connection_created = NULL, .msg_process = st_ipc_dispatch, .connection_closed = st_ipc_closed, .connection_destroyed = st_ipc_destroy }; /*! * \internal * \brief Callback for peer status changes * * \param[in] type What changed * \param[in] node What peer had the change * \param[in] data Previous value of what changed */ static void st_peer_update_callback(enum crm_status_type type, crm_node_t * node, const void *data) { if ((type != crm_status_processes) && !pcmk_is_set(node->flags, crm_remote_node)) { /* * This is a hack until we can send to a nodeid and/or we fix node name lookups * These messages are ignored in stonith_peer_callback() */ xmlNode *query = create_xml_node(NULL, "stonith_command"); crm_xml_add(query, F_XML_TAGNAME, "stonith_command"); crm_xml_add(query, F_TYPE, T_STONITH_NG); crm_xml_add(query, F_STONITH_OPERATION, "poke"); crm_debug("Broadcasting our uname because of node %u", node->id); send_cluster_message(NULL, crm_msg_stonith_ng, query, FALSE); free_xml(query); } } int main(int argc, char **argv) { int flag; int lpc = 0; int argerr = 0; int option_index = 0; crm_cluster_t *cluster = NULL; const char *actions[] = { "reboot", "off", "on", "list", "monitor", "status" }; crm_ipc_t *old_instance = NULL; int rc = pcmk_rc_ok; crm_log_preinit(NULL, argc, argv); pcmk__set_cli_options(NULL, "[options]", long_options, "daemon for executing fencing devices in a " "Pacemaker cluster"); while (1) { flag = pcmk__next_cli_option(argc, argv, &option_index, NULL); if (flag == -1) { break; } switch (flag) { case 'V': crm_bump_log_level(argc, argv); break; case 'l': { int rc = pcmk__add_logfile(optarg); if (rc != pcmk_rc_ok) { /* Logging has not yet been initialized, so stderr is * the only way to get information out */ fprintf(stderr, "Logging to %s is disabled: %s\n", optarg, pcmk_rc_str(rc)); } } break; case 's': stand_alone = TRUE; break; case 'c': stand_alone = FALSE; no_cib_connect = TRUE; break; case '$': case '?': pcmk__cli_help(flag, CRM_EX_OK); break; default: ++argerr; break; } } if (argc - optind == 1 && pcmk__str_eq("metadata", argv[optind], pcmk__str_casei)) { printf("\n"); printf("\n"); printf(" 1.0\n"); printf(" Instance attributes available for all \"stonith\"-class resources" " and used by Pacemaker's fence daemon, formerly known as stonithd\n"); printf(" Instance attributes available for all \"stonith\"-class resources\n"); printf(" \n"); #if 0 // priority is not implemented yet printf(" \n"); printf(" Devices that are not in a topology " "are tried in order of highest to lowest integer priority\n"); printf(" \n"); printf(" \n"); #endif printf(" \n", PCMK_STONITH_HOST_ARGUMENT); printf (" Advanced use only: An alternate parameter to supply instead of 'port'\n"); printf (" Some devices do not support the standard 'port' parameter or may provide additional ones.\n" "Use this to specify an alternate, device-specific, parameter that should indicate the machine to be fenced.\n" "A value of 'none' can be used to tell the cluster not to supply any additional parameters.\n" " \n"); printf(" \n"); printf(" \n"); printf(" \n", PCMK_STONITH_HOST_MAP); printf (" A mapping of host names to ports numbers for devices that do not support host names.\n"); printf (" Eg. node1:1;node2:2,3 would tell the cluster to use port 1 for node1 and ports 2 and 3 for node2\n"); printf(" \n"); printf(" \n"); printf(" \n", PCMK_STONITH_HOST_LIST); printf(" A list of machines controlled by " "this device (Optional unless %s=static-list).\n", PCMK_STONITH_HOST_CHECK); printf(" \n"); printf(" \n"); printf(" \n", PCMK_STONITH_HOST_CHECK); printf (" How to determine which machines are controlled by the device.\n"); printf(" Allowed values: dynamic-list " "(query the device via the 'list' command), static-list " "(check the " PCMK_STONITH_HOST_LIST " attribute), status " "(query the device via the 'status' command), none (assume " "every device can fence every machine)\n"); printf(" \n"); printf(" \n"); printf(" \n", PCMK_STONITH_DELAY_MAX); printf(" Enable a delay of no more than the " "time specified before executing fencing actions. Pacemaker " "derives the overall delay by taking the value of " PCMK_STONITH_DELAY_BASE " and adding a random delay value such " "that the sum is kept below this maximum.\n"); printf(" This prevents double fencing when " "using slow devices such as sbd.\nUse this to enable a random " "delay for fencing actions.\nThe overall delay is derived from " "this random delay value adding a static delay so that the sum " "is kept below the maximum delay.\n"); printf(" \n"); printf(" \n"); printf(" \n", PCMK_STONITH_DELAY_BASE); printf(" Enable a base delay for " "fencing actions and specify base delay value.\n"); printf(" This prevents double fencing when " "different delays are configured on the nodes.\nUse this to " "enable a static delay for fencing actions.\nThe overall delay " "is derived from a random delay value adding this static delay " "so that the sum is kept below the maximum delay.\nSet to eg. " "node1:1s;node2:5 to set different value per node.\n"); printf(" \n"); printf(" \n"); printf(" \n", PCMK_STONITH_ACTION_LIMIT); printf (" The maximum number of actions can be performed in parallel on this device\n"); printf (" Cluster property concurrent-fencing=true needs to be configured first.\n" "Then use this to specify the maximum number of actions can be performed in parallel on this device. -1 is unlimited.\n"); printf(" \n"); printf(" \n"); for (lpc = 0; lpc < PCMK__NELEM(actions); lpc++) { printf(" \n", actions[lpc]); printf (" Advanced use only: An alternate command to run instead of '%s'\n", actions[lpc]); printf (" Some devices do not support the standard commands or may provide additional ones.\n" "Use this to specify an alternate, device-specific, command that implements the '%s' action.\n", actions[lpc]); printf(" \n", actions[lpc]); printf(" \n"); printf(" \n", actions[lpc]); printf (" Advanced use only: Specify an alternate timeout to use for %s actions instead of stonith-timeout\n", actions[lpc]); printf (" Some devices need much more/less time to complete than normal.\n" "Use this to specify an alternate, device-specific, timeout for '%s' actions.\n", actions[lpc]); printf(" \n"); printf(" \n"); printf(" \n", actions[lpc]); printf (" Advanced use only: The maximum number of times to retry the '%s' command within the timeout period\n", actions[lpc]); printf(" Some devices do not support multiple connections." " Operations may 'fail' if the device is busy with another task so Pacemaker will automatically retry the operation, if there is time remaining." " Use this option to alter the number of times Pacemaker retries '%s' actions before giving up." "\n", actions[lpc]); printf(" \n"); printf(" \n"); } printf(" \n"); printf("\n"); return CRM_EX_OK; } if (optind != argc) { ++argerr; } if (argerr) { pcmk__cli_help('?', CRM_EX_USAGE); } crm_log_init(NULL, LOG_INFO, TRUE, FALSE, argc, argv, FALSE); crm_notice("Starting Pacemaker fencer"); old_instance = crm_ipc_new("stonith-ng", 0); if (crm_ipc_connect(old_instance)) { /* IPC end-point already up */ crm_ipc_close(old_instance); crm_ipc_destroy(old_instance); crm_err("pacemaker-fenced is already active, aborting startup"); crm_exit(CRM_EX_OK); } else { /* not up or not authentic, we'll proceed either way */ crm_ipc_destroy(old_instance); old_instance = NULL; } mainloop_add_signal(SIGTERM, stonith_shutdown); crm_peer_init(); fenced_data_set = pe_new_working_set(); CRM_ASSERT(fenced_data_set != NULL); pe__set_working_set_flags(fenced_data_set, pe_flag_no_counts|pe_flag_no_compat); pe__set_working_set_flags(fenced_data_set, pe_flag_show_utilization); cluster = calloc(1, sizeof(crm_cluster_t)); CRM_ASSERT(cluster != NULL); if (stand_alone == FALSE) { if (is_corosync_cluster()) { #if SUPPORT_COROSYNC cluster->destroy = stonith_peer_cs_destroy; cluster->cpg.cpg_deliver_fn = stonith_peer_ais_callback; cluster->cpg.cpg_confchg_fn = pcmk_cpg_membership; #endif } crm_set_status_callback(&st_peer_update_callback); if (crm_cluster_connect(cluster) == FALSE) { crm_crit("Cannot sign in to the cluster... terminating"); crm_exit(CRM_EX_FATAL); } stonith_our_uname = strdup(cluster->uname); if (no_cib_connect == FALSE) { setup_cib(); } } else { stonith_our_uname = strdup("localhost"); } init_device_list(); init_topology_list(); pcmk__serve_fenced_ipc(&ipcs, &ipc_callbacks); pcmk__register_formats(NULL, formats); rc = pcmk__output_new(&out, "log", NULL, argv); if ((rc != pcmk_rc_ok) || (out == NULL)) { crm_err("Can't log resource details due to internal error: %s\n", pcmk_rc_str(rc)); crm_exit(CRM_EX_FATAL); } pe__register_messages(out); pcmk__register_lib_messages(out); pcmk__output_set_log_level(out, LOG_TRACE); fenced_data_set->priv = out; /* Create the mainloop and run it... */ mainloop = g_main_loop_new(NULL, FALSE); crm_notice("Pacemaker fencer successfully started and accepting connections"); g_main_loop_run(mainloop); stonith_cleanup(); free(cluster->uuid); free(cluster->uname); free(cluster); pe_free_working_set(fenced_data_set); out->finish(out, CRM_EX_OK, true, NULL); pcmk__output_free(out); pcmk__unregister_formats(); crm_exit(CRM_EX_OK); } diff --git a/daemons/fenced/pacemaker-fenced.h b/daemons/fenced/pacemaker-fenced.h index 5162ada75d..cf114fb979 100644 --- a/daemons/fenced/pacemaker-fenced.h +++ b/daemons/fenced/pacemaker-fenced.h @@ -1,279 +1,280 @@ /* * Copyright 2009-2021 the Pacemaker project contributors * * This source code is licensed under the GNU General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include // uint32_t, uint64_t #include /*! * \internal * \brief Check whether target has already been fenced recently * * \param[in] tolerance Number of seconds to look back in time * \param[in] target Name of node to search for * \param[in] action Action we want to match * * \return TRUE if an equivalent fencing operation took place in the last * \p tolerance seconds, FALSE otherwise */ gboolean stonith_check_fence_tolerance(int tolerance, const char *target, const char *action); typedef struct stonith_device_s { char *id; char *agent; char *namespace; /*! list of actions that must execute on the target node. Used for unfencing */ char *on_target_actions; GList *targets; time_t targets_age; gboolean has_attr_map; /* should nodeid parameter for victim be included in agent arguments */ gboolean include_nodeid; /* whether the cluster should automatically unfence nodes with the device */ gboolean automatic_unfencing; guint priority; uint32_t flags; // Group of enum st_device_flags GHashTable *params; GHashTable *aliases; GList *pending_ops; mainloop_timer_t *timer; crm_trigger_t *work; xmlNode *agent_metadata; /*! A verified device is one that has contacted the * agent successfully to perform a monitor operation */ gboolean verified; gboolean cib_registered; gboolean api_registered; gboolean dirty; } stonith_device_t; /* These values are used to index certain arrays by "phase". Usually an * operation has only one "phase", so phase is always zero. However, some * reboots are remapped to "off" then "on", in which case "reboot" will be * phase 0, "off" will be phase 1 and "on" will be phase 2. */ enum st_remap_phase { st_phase_requested = 0, st_phase_off = 1, st_phase_on = 2, st_phase_max = 3 }; typedef struct remote_fencing_op_s { /* The unique id associated with this operation */ char *id; /*! The node this operation will fence */ char *target; /*! The fencing action to perform on the target. (reboot, on, off) */ char *action; /*! When was the fencing action recorded (seconds since epoch) */ time_t created; /*! Marks if the final notifications have been sent to local stonith clients. */ gboolean notify_sent; /*! The number of query replies received */ guint replies; /*! The number of query replies expected */ guint replies_expected; /*! Does this node own control of this operation */ gboolean owner; /*! After query is complete, This the high level timer that expires the entire operation */ guint op_timer_total; /*! This timer expires the current fencing request. Many fencing * requests may exist in a single operation */ guint op_timer_one; /*! This timer expires the query request sent out to determine * what nodes are contain what devices, and who those devices can fence */ guint query_timer; /*! This is the default timeout to use for each fencing device if no * custom timeout is received in the query. */ gint base_timeout; /*! This is the calculated total timeout an operation can take before * expiring. This is calculated by adding together all the timeout * values associated with the devices this fencing operation may call */ gint total_timeout; /*! Requested fencing delay. * Value -1 means disable any static/random fencing delays. */ int delay; /*! Delegate is the node being asked to perform a fencing action * on behalf of the node that owns the remote operation. Some operations * will involve multiple delegates. This value represents the final delegate * that is used. */ char *delegate; /*! The point at which the remote operation completed */ time_t completed; //! Group of enum stonith_call_options associated with this operation uint32_t call_options; /*! The current state of the remote operation. This indicates * what stage the op is in, query, exec, done, duplicate, failed. */ enum op_state state; /*! The node that owns the remote operation */ char *originator; /*! The local client id that initiated the fencing request */ char *client_id; /*! The client's call_id that initiated the fencing request */ int client_callid; /*! The name of client that initiated the fencing request */ char *client_name; /*! List of the received query results for all the nodes in the cpg group */ GList *query_results; /*! The original request that initiated the remote stonith operation */ xmlNode *request; /*! The current topology level being executed */ guint level; /*! The current operation phase being executed */ enum st_remap_phase phase; /*! Devices with automatic unfencing (always run if "on" requested, never if remapped) */ GList *automatic_list; /*! List of all devices at the currently executing topology level */ GList *devices_list; /*! Current entry in the topology device list */ GList *devices; /*! List of duplicate operations attached to this operation. Once this operation * completes, the duplicate operations will be closed out as well. */ GList *duplicates; /*! The point at which the remote operation completed(nsec) */ long long completed_nsec; } remote_fencing_op_t; /*! * \internal * \brief Broadcast the result of an operation to the peers. * \param op, Operation whose result should be broadcast * \param rc, Result of the operation */ void stonith_bcast_result_to_peers(remote_fencing_op_t * op, int rc, gboolean op_merged); // Fencer-specific client flags enum st_client_flags { st_callback_unknown = UINT64_C(0), st_callback_notify_fence = (UINT64_C(1) << 0), st_callback_device_add = (UINT64_C(1) << 2), st_callback_device_del = (UINT64_C(1) << 4), st_callback_notify_history = (UINT64_C(1) << 5), st_callback_notify_history_synced = (UINT64_C(1) << 6) }; /* * Complex fencing requirements are specified via fencing topologies. * A topology consists of levels; each level is a list of fencing devices. * Topologies are stored in a hash table by node name. When a node needs to be * fenced, if it has an entry in the topology table, the levels are tried * sequentially, and the devices in each level are tried sequentially. * Fencing is considered successful as soon as any level succeeds; * a level is considered successful if all its devices succeed. * Essentially, all devices at a given level are "and-ed" and the * levels are "or-ed". * * This structure is used for the topology table entries. * Topology levels start from 1, so levels[0] is unused and always NULL. */ typedef struct stonith_topology_s { int kind; /*! Node name regex or attribute name=value for which topology applies */ char *target; char *target_value; char *target_pattern; char *target_attribute; /*! Names of fencing devices at each topology level */ GList *levels[ST_LEVEL_MAX]; } stonith_topology_t; void init_device_list(void); void free_device_list(void); void init_topology_list(void); void free_topology_list(void); void free_stonith_remote_op_list(void); void init_stonith_remote_op_hash_table(GHashTable **table); void free_metadata_cache(void); uint64_t get_stonith_flag(const char *name); void stonith_command(pcmk__client_t *client, uint32_t id, uint32_t flags, xmlNode *op_request, const char *remote_peer); int stonith_device_register(xmlNode * msg, const char **desc, gboolean from_cib); void stonith_device_remove(const char *id, bool from_cib); char *stonith_level_key(xmlNode * msg, int mode); int stonith_level_kind(xmlNode * msg); -int stonith_level_register(xmlNode * msg, char **desc); +void fenced_register_level(xmlNode *msg, char **desc, + pcmk__action_result_t *result); int stonith_level_remove(xmlNode * msg, char **desc); stonith_topology_t *find_topology_for_host(const char *host); void do_local_reply(xmlNode * notify_src, const char *client_id, gboolean sync_reply, gboolean from_peer); xmlNode *stonith_construct_reply(xmlNode * request, const char *output, xmlNode * data, int rc); void do_stonith_async_timeout_update(const char *client, const char *call_id, int timeout); void do_stonith_notify(const char *type, int result, xmlNode *data); void do_stonith_notify_device(const char *op, int rc, const char *desc); void do_stonith_notify_level(const char *op, int rc, const char *desc); remote_fencing_op_t *initiate_remote_stonith_op(pcmk__client_t *client, xmlNode *request, gboolean manual_ack); void fenced_process_fencing_reply(xmlNode *msg); int process_remote_stonith_query(xmlNode * msg); void *create_remote_stonith_op(const char *client, xmlNode * request, gboolean peer); void stonith_fence_history(xmlNode *msg, xmlNode **output, const char *remote_peer, int options); void stonith_fence_history_trim(void); bool fencing_peer_active(crm_node_t *peer); void set_fencing_completed(remote_fencing_op_t * op); int fenced_handle_manual_confirmation(pcmk__client_t *client, xmlNode *msg); gboolean node_has_attr(const char *node, const char *name, const char *value); gboolean node_does_watchdog_fencing(const char *node); static inline void fenced_set_protocol_error(pcmk__action_result_t *result) { pcmk__set_result(result, CRM_EX_PROTOCOL, PCMK_EXEC_INVALID, "Fencer API request missing required information (bug?)"); } extern char *stonith_our_uname; extern gboolean stand_alone; extern GHashTable *device_list; extern GHashTable *topology; extern long stonith_watchdog_timeout_ms; extern GList *stonith_watchdog_targets; extern GHashTable *stonith_remote_op_list; diff --git a/lib/fencing/st_actions.c b/lib/fencing/st_actions.c index 7eaa8b0f2b..37fa849847 100644 --- a/lib/fencing/st_actions.c +++ b/lib/fencing/st_actions.c @@ -1,689 +1,690 @@ /* * Copyright 2004-2021 the Pacemaker project contributors * * The version control history for this file may have further details. * * This source code is licensed under the GNU Lesser General Public License * version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include "fencing_private.h" struct stonith_action_s { /*! user defined data */ char *agent; char *action; char *victim; GHashTable *args; int timeout; int async; void *userdata; void (*done_cb) (int pid, const pcmk__action_result_t *result, void *user_data); void (*fork_cb) (int pid, void *user_data); svc_action_t *svc_action; /*! internal timing information */ time_t initial_start_time; int tries; int remaining_timeout; int max_retries; int pid; pcmk__action_result_t result; }; static int internal_stonith_action_execute(stonith_action_t *action); static void log_action(stonith_action_t *action, pid_t pid); /*! * \internal * \brief Set an action's result based on services library result * * \param[in] action Fence action to set result for * \param[in] svc_action Service action to get result from */ static void set_result_from_svc_action(stonith_action_t *action, svc_action_t *svc_action) { pcmk__set_result(&(action->result), svc_action->rc, svc_action->status, services__exit_reason(svc_action)); pcmk__set_result_output(&(action->result), services__grab_stdout(svc_action), services__grab_stderr(svc_action)); } static void log_action(stonith_action_t *action, pid_t pid) { /* The services library has already logged the output at info or debug * level, so just raise to warning for stderr. */ if (action->result.action_stderr != NULL) { /* Logging the whole string confuses syslog when the string is xml */ char *prefix = crm_strdup_printf("%s[%d] stderr:", action->agent, pid); crm_log_output(LOG_WARNING, prefix, action->result.action_stderr); free(prefix); } } static void append_config_arg(gpointer key, gpointer value, gpointer user_data) { /* The fencer will filter "action" out when it registers the device, * but ignore it here in case any external API users don't. * * Also filter out parameters handled directly by Pacemaker. */ if (!pcmk__str_eq(key, STONITH_ATTR_ACTION_OP, pcmk__str_casei) && !pcmk_stonith_param(key) && (strstr(key, CRM_META) == NULL) && !pcmk__str_eq(key, "crm_feature_set", pcmk__str_casei)) { crm_trace("Passing %s=%s with fence action", (const char *) key, (const char *) (value? value : "")); g_hash_table_insert((GHashTable *) user_data, strdup(key), strdup(value? value : "")); } } static GHashTable * make_args(const char *agent, const char *action, const char *victim, uint32_t victim_nodeid, GHashTable * device_args, GHashTable * port_map, const char *host_arg) { GHashTable *arg_list = NULL; const char *value = NULL; CRM_CHECK(action != NULL, return NULL); arg_list = pcmk__strkey_table(free, free); // Add action to arguments (using an alias if requested) if (device_args) { char buffer[512]; snprintf(buffer, sizeof(buffer), "pcmk_%s_action", action); value = g_hash_table_lookup(device_args, buffer); if (value) { crm_debug("Substituting '%s' for fence action %s targeting %s", value, action, victim); action = value; } } g_hash_table_insert(arg_list, strdup(STONITH_ATTR_ACTION_OP), strdup(action)); /* If this is a fencing operation against another node, add more standard * arguments. */ if (victim && device_args) { const char *param = NULL; /* Always pass the target's name, per * https://github.com/ClusterLabs/fence-agents/blob/master/doc/FenceAgentAPI.md */ g_hash_table_insert(arg_list, strdup("nodename"), strdup(victim)); // If the target's node ID was specified, pass it, too if (victim_nodeid) { char *nodeid = crm_strdup_printf("%" PRIu32, victim_nodeid); // cts-fencing looks for this log message crm_info("Passing '%s' as nodeid with fence action '%s' targeting %s", nodeid, action, victim); g_hash_table_insert(arg_list, strdup("nodeid"), nodeid); } // Check whether target must be specified in some other way param = g_hash_table_lookup(device_args, PCMK_STONITH_HOST_ARGUMENT); if (!pcmk__str_eq(agent, "fence_legacy", pcmk__str_none) && !pcmk__str_eq(param, "none", pcmk__str_casei)) { if (param == NULL) { /* Use the caller's default for pcmk_host_argument, or "port" if * none was given */ param = (host_arg == NULL)? "port" : host_arg; } value = g_hash_table_lookup(device_args, param); if (pcmk__str_eq(value, "dynamic", pcmk__str_casei|pcmk__str_null_matches)) { /* If the host argument was "dynamic" or not explicitly specified, * add it with the target */ const char *alias = NULL; if (port_map) { alias = g_hash_table_lookup(port_map, victim); } if (alias == NULL) { alias = victim; } crm_debug("Passing %s='%s' with fence action %s targeting %s", param, alias, action, victim); g_hash_table_insert(arg_list, strdup(param), strdup(alias)); } } } if (device_args) { g_hash_table_foreach(device_args, append_config_arg, arg_list); } return arg_list; } /*! * \internal * \brief Free all memory used by a stonith action * * \param[in,out] action Action to free */ void stonith__destroy_action(stonith_action_t *action) { if (action) { free(action->agent); if (action->args) { g_hash_table_destroy(action->args); } free(action->action); free(action->victim); if (action->svc_action) { services_action_free(action->svc_action); } pcmk__reset_result(&(action->result)); free(action); } } /*! * \internal * \brief Get the result of an executed stonith action * * \param[in] action Executed action * * \return Pointer to action's result (or NULL if \p action is NULL) */ pcmk__action_result_t * stonith__action_result(stonith_action_t *action) { return (action == NULL)? NULL : &(action->result); } #define FAILURE_MAX_RETRIES 2 stonith_action_t * stonith_action_create(const char *agent, const char *_action, const char *victim, uint32_t victim_nodeid, int timeout, GHashTable * device_args, GHashTable * port_map, const char *host_arg) { stonith_action_t *action; action = calloc(1, sizeof(stonith_action_t)); action->args = make_args(agent, _action, victim, victim_nodeid, device_args, port_map, host_arg); crm_debug("Preparing '%s' action for %s using agent %s", _action, (victim? victim : "no target"), agent); action->agent = strdup(agent); action->action = strdup(_action); if (victim) { action->victim = strdup(victim); } action->timeout = action->remaining_timeout = timeout; action->max_retries = FAILURE_MAX_RETRIES; pcmk__set_result(&(action->result), PCMK_OCF_UNKNOWN, PCMK_EXEC_UNKNOWN, "Initialization bug in fencing library"); if (device_args) { char buffer[512]; const char *value = NULL; snprintf(buffer, sizeof(buffer), "pcmk_%s_retries", _action); value = g_hash_table_lookup(device_args, buffer); if (value) { action->max_retries = atoi(value); } } return action; } static gboolean update_remaining_timeout(stonith_action_t * action) { int diff = time(NULL) - action->initial_start_time; if (action->tries >= action->max_retries) { crm_info("Attempted to execute agent %s (%s) the maximum number of times (%d) allowed", action->agent, action->action, action->max_retries); action->remaining_timeout = 0; } else if ((action->result.execution_status != PCMK_EXEC_TIMEOUT) && (diff < (action->timeout * 0.7))) { /* only set remaining timeout period if there is 30% * or greater of the original timeout period left */ action->remaining_timeout = action->timeout - diff; } else { action->remaining_timeout = 0; } return action->remaining_timeout ? TRUE : FALSE; } /*! * \internal * \brief Map a fencing action result to a standard return code * * \param[in] result Fencing action result to map * * \return Standard Pacemaker return code that best corresponds to \p result */ int stonith__result2rc(const pcmk__action_result_t *result) { if (pcmk__result_ok(result)) { return pcmk_rc_ok; } switch (result->execution_status) { case PCMK_EXEC_PENDING: return EINPROGRESS; case PCMK_EXEC_CANCELLED: return ECANCELED; case PCMK_EXEC_TIMEOUT: return ETIME; case PCMK_EXEC_NOT_INSTALLED: return ENOENT; case PCMK_EXEC_NOT_SUPPORTED: return EOPNOTSUPP; case PCMK_EXEC_NOT_CONNECTED: return ENOTCONN; case PCMK_EXEC_NO_FENCE_DEVICE: return ENODEV; case PCMK_EXEC_NO_SECRETS: return EACCES; /* For the fencing API, PCMK_EXEC_INVALID is used with fencer API * operations that don't involve executing an agent (for example, * registering devices). This allows us to use the CRM_EX_* codes in the * exit status for finer-grained responses. */ case PCMK_EXEC_INVALID: switch (result->exit_status) { + case CRM_EX_INVALID_PARAM: return EINVAL; case CRM_EX_INSUFFICIENT_PRIV: return EACCES; case CRM_EX_PROTOCOL: return EPROTO; /* CRM_EX_EXPIRED is used for orphaned fencing operations left * over from a previous instance of the fencer. For API backward * compatibility, this is mapped to the previously used code for * this case, EHOSTUNREACH. */ case CRM_EX_EXPIRED: return EHOSTUNREACH; default: break; } break; default: break; } // Try to provide useful error code based on result's error output if (result->action_stderr == NULL) { return ENODATA; } else if (strcasestr(result->action_stderr, "timed out") || strcasestr(result->action_stderr, "timeout")) { return ETIME; } else if (strcasestr(result->action_stderr, "unrecognised action") || strcasestr(result->action_stderr, "unrecognized action") || strcasestr(result->action_stderr, "unsupported action")) { return EOPNOTSUPP; } // Oh well, we tried return pcmk_rc_error; } /*! * \internal * \brief Determine execution status equivalent of legacy fencer return code * * Fence action notifications, and fence action callbacks from older fencers * (<=2.1.2) in a rolling upgrade, will have only a legacy return code. Map this * to an execution status as best as possible (essentially, the inverse of * stonith__result2rc()). * * \param[in] rc Legacy return code from fencer * * \return Execution status best corresponding to \p rc */ int stonith__legacy2status(int rc) { if (rc >= 0) { return PCMK_EXEC_DONE; } switch (-rc) { case EACCES: return PCMK_EXEC_NO_SECRETS; case ECANCELED: return PCMK_EXEC_CANCELLED; case EHOSTUNREACH: return PCMK_EXEC_INVALID; case EINPROGRESS: return PCMK_EXEC_PENDING; case ENODEV: return PCMK_EXEC_NO_FENCE_DEVICE; case ENOENT: return PCMK_EXEC_NOT_INSTALLED; case ENOTCONN: return PCMK_EXEC_NOT_CONNECTED; case EOPNOTSUPP: return PCMK_EXEC_NOT_SUPPORTED; case EPROTO: return PCMK_EXEC_INVALID; case EPROTONOSUPPORT: return PCMK_EXEC_NOT_SUPPORTED; case ETIME: return PCMK_EXEC_TIMEOUT; case ETIMEDOUT: return PCMK_EXEC_TIMEOUT; default: return PCMK_EXEC_ERROR; } } /*! * \internal * \brief Add a fencing result to an XML element as attributes * * \param[in] xml XML element to add result to * \param[in] result Fencing result to add (assume success if NULL) */ void stonith__xe_set_result(xmlNode *xml, const pcmk__action_result_t *result) { int exit_status = CRM_EX_OK; enum pcmk_exec_status execution_status = PCMK_EXEC_DONE; const char *exit_reason = NULL; const char *action_stdout = NULL; int rc = pcmk_ok; CRM_CHECK(xml != NULL, return); if (result != NULL) { exit_status = result->exit_status; execution_status = result->execution_status; exit_reason = result->exit_reason; action_stdout = result->action_stdout; rc = pcmk_rc2legacy(stonith__result2rc(result)); } crm_xml_add_int(xml, XML_LRM_ATTR_OPSTATUS, (int) execution_status); crm_xml_add_int(xml, XML_LRM_ATTR_RC, exit_status); crm_xml_add(xml, XML_LRM_ATTR_EXIT_REASON, exit_reason); crm_xml_add(xml, F_STONITH_OUTPUT, action_stdout); /* @COMPAT Peers in rolling upgrades, Pacemaker Remote nodes, and external * code that use libstonithd <=2.1.2 don't check for the full result, and * need a legacy return code instead. */ crm_xml_add_int(xml, F_STONITH_RC, rc); } /*! * \internal * \brief Find a fencing result beneath an XML element * * \param[in] xml XML element to search * * \return \p xml or descendent of it that contains a fencing result, else NULL */ xmlNode * stonith__find_xe_with_result(xmlNode *xml) { xmlNode *match = get_xpath_object("//@" XML_LRM_ATTR_RC, xml, LOG_NEVER); if (match == NULL) { /* @COMPAT Peers <=2.1.2 in a rolling upgrade provide only a legacy * return code, not a full result, so check for that. */ match = get_xpath_object("//@" F_STONITH_RC, xml, LOG_ERR); } return match; } /*! * \internal * \brief Get a fencing result from an XML element's attributes * * \param[in] xml XML element with fencing result * \param[out] result Where to store fencing result */ void stonith__xe_get_result(xmlNode *xml, pcmk__action_result_t *result) { int exit_status = CRM_EX_OK; int execution_status = PCMK_EXEC_DONE; const char *exit_reason = NULL; char *action_stdout = NULL; CRM_CHECK((xml != NULL) && (result != NULL), return); exit_reason = crm_element_value(xml, XML_LRM_ATTR_EXIT_REASON); action_stdout = crm_element_value_copy(xml, F_STONITH_OUTPUT); // A result must include an exit status and execution status if ((crm_element_value_int(xml, XML_LRM_ATTR_RC, &exit_status) < 0) || (crm_element_value_int(xml, XML_LRM_ATTR_OPSTATUS, &execution_status) < 0)) { int rc = pcmk_ok; exit_status = CRM_EX_ERROR; /* @COMPAT Peers <=2.1.2 in rolling upgrades provide only a legacy * return code, not a full result, so check for that. */ if (crm_element_value_int(xml, F_STONITH_RC, &rc) == 0) { if ((rc == pcmk_ok) || (rc == -EINPROGRESS)) { exit_status = CRM_EX_OK; } execution_status = stonith__legacy2status(rc); exit_reason = pcmk_strerror(rc); } else { execution_status = PCMK_EXEC_ERROR; exit_reason = "Fencer reply contained neither a full result " "nor a legacy return code (bug?)"; } } pcmk__set_result(result, exit_status, execution_status, exit_reason); pcmk__set_result_output(result, action_stdout, NULL); } static void stonith_action_async_done(svc_action_t *svc_action) { stonith_action_t *action = (stonith_action_t *) svc_action->cb_data; set_result_from_svc_action(action, svc_action); svc_action->params = NULL; crm_debug("Child process %d performing action '%s' exited with rc %d", action->pid, action->action, svc_action->rc); log_action(action, action->pid); if (!pcmk__result_ok(&(action->result)) && update_remaining_timeout(action)) { int rc = internal_stonith_action_execute(action); if (rc == pcmk_ok) { return; } } if (action->done_cb) { action->done_cb(action->pid, &(action->result), action->userdata); } action->svc_action = NULL; // don't remove our caller stonith__destroy_action(action); } static void stonith_action_async_forked(svc_action_t *svc_action) { stonith_action_t *action = (stonith_action_t *) svc_action->cb_data; action->pid = svc_action->pid; action->svc_action = svc_action; if (action->fork_cb) { (action->fork_cb) (svc_action->pid, action->userdata); } crm_trace("Child process %d performing action '%s' successfully forked", action->pid, action->action); } static int internal_stonith_action_execute(stonith_action_t * action) { int rc = -EPROTO; int is_retry = 0; svc_action_t *svc_action = NULL; static int stonith_sequence = 0; char *buffer = NULL; CRM_CHECK(action != NULL, return -EINVAL); if ((action->action == NULL) || (action->args == NULL) || (action->agent == NULL)) { pcmk__set_result(&(action->result), PCMK_OCF_UNKNOWN_ERROR, PCMK_EXEC_ERROR_FATAL, "Bug in fencing library"); return -EINVAL; } if (!action->tries) { action->initial_start_time = time(NULL); } action->tries++; if (action->tries > 1) { crm_info("Attempt %d to execute %s (%s). remaining timeout is %d", action->tries, action->agent, action->action, action->remaining_timeout); is_retry = 1; } buffer = crm_strdup_printf(PCMK__FENCE_BINDIR "/%s", basename(action->agent)); svc_action = services_action_create_generic(buffer, NULL); free(buffer); if (svc_action->rc != PCMK_OCF_UNKNOWN) { set_result_from_svc_action(action, svc_action); services_action_free(svc_action); return -E2BIG; } svc_action->timeout = 1000 * action->remaining_timeout; svc_action->standard = strdup(PCMK_RESOURCE_CLASS_STONITH); svc_action->id = crm_strdup_printf("%s_%s_%d", basename(action->agent), action->action, action->tries); svc_action->agent = strdup(action->agent); svc_action->sequence = stonith_sequence++; svc_action->params = action->args; svc_action->cb_data = (void *) action; svc_action->flags = pcmk__set_flags_as(__func__, __LINE__, LOG_TRACE, "Action", svc_action->id, svc_action->flags, SVC_ACTION_NON_BLOCKED, "SVC_ACTION_NON_BLOCKED"); /* keep retries from executing out of control and free previous results */ if (is_retry) { pcmk__reset_result(&(action->result)); sleep(1); } if (action->async) { /* async */ if (services_action_async_fork_notify(svc_action, &stonith_action_async_done, &stonith_action_async_forked)) { pcmk__set_result(&(action->result), PCMK_OCF_UNKNOWN, PCMK_EXEC_PENDING, NULL); return pcmk_ok; } } else if (services_action_sync(svc_action)) { // sync success rc = pcmk_ok; } else { // sync failure rc = -ECONNABORTED; } set_result_from_svc_action(action, svc_action); svc_action->params = NULL; services_action_free(svc_action); return rc; } /*! * \internal * \brief Kick off execution of an async stonith action * * \param[in,out] action Action to be executed * \param[in,out] userdata Datapointer to be passed to callbacks * \param[in] done Callback to notify action has failed/succeeded * \param[in] fork_callback Callback to notify successful fork of child * * \return pcmk_ok if ownership of action has been taken, -errno otherwise */ int stonith_action_execute_async(stonith_action_t * action, void *userdata, void (*done) (int pid, const pcmk__action_result_t *result, void *user_data), void (*fork_cb) (int pid, void *user_data)) { if (!action) { return -EINVAL; } action->userdata = userdata; action->done_cb = done; action->fork_cb = fork_cb; action->async = 1; return internal_stonith_action_execute(action); } /*! * \internal * \brief Execute a stonith action * * \param[in,out] action Action to execute * * \return pcmk_ok on success, -errno otherwise */ int stonith__execute(stonith_action_t *action) { int rc = pcmk_ok; CRM_CHECK(action != NULL, return -EINVAL); // Keep trying until success, max retries, or timeout do { rc = internal_stonith_action_execute(action); } while ((rc != pcmk_ok) && update_remaining_timeout(action)); return rc; }