diff --git a/daemons/controld/controld_fencing.c b/daemons/controld/controld_fencing.c index a848e05488..cdfa08192b 100644 --- a/daemons/controld/controld_fencing.c +++ b/daemons/controld/controld_fencing.c @@ -1,1106 +1,1106 @@ /* * Copyright 2004-2023 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 static void tengine_stonith_history_synced(stonith_t *st, stonith_event_t *st_event); /* * stonith failure counting * * We don't want to get stuck in a permanent fencing loop. Keep track of the * number of fencing failures for each target node, and the most we'll restart a * transition for. */ struct st_fail_rec { int count; }; static bool fence_reaction_panic = false; static unsigned long int stonith_max_attempts = 10; static GHashTable *stonith_failures = NULL; /*! * \internal * \brief Update max fencing attempts before giving up * * \param[in] value New max fencing attempts */ static void update_stonith_max_attempts(const char *value) { stonith_max_attempts = char2score(value); if (stonith_max_attempts < 1UL) { stonith_max_attempts = 10UL; } } /*! * \internal * \brief Configure reaction to notification of local node being fenced * * \param[in] reaction_s Reaction type */ static void set_fence_reaction(const char *reaction_s) { if (pcmk__str_eq(reaction_s, "panic", pcmk__str_casei)) { fence_reaction_panic = true; } else { if (!pcmk__str_eq(reaction_s, "stop", pcmk__str_casei)) { crm_warn("Invalid value '%s' for %s, using 'stop'", reaction_s, XML_CONFIG_ATTR_FENCE_REACTION); } fence_reaction_panic = false; } } /*! * \internal * \brief Configure fencing options based on the CIB * * \param[in,out] options Name/value pairs for configured options */ void controld_configure_fencing(GHashTable *options) { const char *value = NULL; value = g_hash_table_lookup(options, XML_CONFIG_ATTR_FENCE_REACTION); set_fence_reaction(value); value = g_hash_table_lookup(options, "stonith-max-attempts"); update_stonith_max_attempts(value); } static gboolean too_many_st_failures(const char *target) { GHashTableIter iter; const char *key = NULL; struct st_fail_rec *value = NULL; if (stonith_failures == NULL) { return FALSE; } if (target == NULL) { g_hash_table_iter_init(&iter, stonith_failures); while (g_hash_table_iter_next(&iter, (gpointer *) &key, (gpointer *) &value)) { if (value->count >= stonith_max_attempts) { target = (const char*)key; goto too_many; } } } else { value = g_hash_table_lookup(stonith_failures, target); if ((value != NULL) && (value->count >= stonith_max_attempts)) { goto too_many; } } return FALSE; too_many: crm_warn("Too many failures (%d) to fence %s, giving up", value->count, target); return TRUE; } /*! * \internal * \brief Reset a stonith fail count * * \param[in] target Name of node to reset, or NULL for all */ void st_fail_count_reset(const char *target) { if (stonith_failures == NULL) { return; } if (target) { struct st_fail_rec *rec = NULL; rec = g_hash_table_lookup(stonith_failures, target); if (rec) { rec->count = 0; } } else { GHashTableIter iter; const char *key = NULL; struct st_fail_rec *rec = NULL; g_hash_table_iter_init(&iter, stonith_failures); while (g_hash_table_iter_next(&iter, (gpointer *) &key, (gpointer *) &rec)) { rec->count = 0; } } } static void st_fail_count_increment(const char *target) { struct st_fail_rec *rec = NULL; if (stonith_failures == NULL) { stonith_failures = pcmk__strkey_table(free, free); } rec = g_hash_table_lookup(stonith_failures, target); if (rec) { rec->count++; } else { rec = malloc(sizeof(struct st_fail_rec)); if(rec == NULL) { return; } rec->count = 1; g_hash_table_insert(stonith_failures, strdup(target), rec); } } /* end stonith fail count functions */ static void cib_fencing_updated(xmlNode *msg, int call_id, int rc, xmlNode *output, void *user_data) { if (rc < pcmk_ok) { crm_err("Fencing update %d for %s: failed - %s (%d)", call_id, (char *)user_data, pcmk_strerror(rc), rc); crm_log_xml_warn(msg, "Failed update"); abort_transition(INFINITY, pcmk__graph_shutdown, "CIB update failed", NULL); } else { crm_info("Fencing update %d for %s: complete", call_id, (char *)user_data); } } static void send_stonith_update(pcmk__graph_action_t *action, const char *target, const char *uuid) { int rc = pcmk_ok; crm_node_t *peer = NULL; /* We (usually) rely on the membership layer to do node_update_cluster, * and the peer status callback to do node_update_peer, because the node * might have already rejoined before we get the stonith result here. */ int flags = node_update_join | node_update_expected; /* zero out the node-status & remove all LRM status info */ xmlNode *node_state = NULL; CRM_CHECK(target != NULL, return); CRM_CHECK(uuid != NULL, return); /* Make sure the membership and join caches are accurate */ peer = crm_get_peer_full(0, target, CRM_GET_PEER_ANY); CRM_CHECK(peer != NULL, return); if (peer->state == NULL) { /* Usually, we rely on the membership layer to update the cluster state * in the CIB. However, if the node has never been seen, do it here, so * the node is not considered unclean. */ flags |= node_update_cluster; } if (peer->uuid == NULL) { crm_info("Recording uuid '%s' for node '%s'", uuid, target); peer->uuid = strdup(uuid); } crmd_peer_down(peer, TRUE); /* Generate a node state update for the CIB */ node_state = create_node_state_update(peer, flags, NULL, __func__); /* we have to mark whether or not remote nodes have already been fenced */ if (peer->flags & crm_remote_node) { char *now_s = pcmk__ttoa(time(NULL)); crm_xml_add(node_state, XML_NODE_IS_FENCED, now_s); free(now_s); } /* Force our known ID */ crm_xml_add(node_state, XML_ATTR_ID, uuid); - rc = controld_globals.cib_conn->cmds->update(controld_globals.cib_conn, + rc = controld_globals.cib_conn->cmds->modify(controld_globals.cib_conn, XML_CIB_TAG_STATUS, node_state, cib_scope_local |cib_can_create); /* Delay processing the trigger until the update completes */ crm_debug("Sending fencing update %d for %s", rc, target); fsa_register_cib_callback(rc, strdup(target), cib_fencing_updated); // Make sure it sticks /* controld_globals.cib_conn->cmds->bump_epoch(controld_globals.cib_conn, * cib_scope_local); */ controld_delete_node_state(peer->uname, controld_section_all, cib_scope_local); free_xml(node_state); return; } /*! * \internal * \brief Abort transition due to stonith failure * * \param[in] abort_action Whether to restart or stop transition * \param[in] target Don't restart if this (NULL for any) has too many failures * \param[in] reason Log this stonith action XML as abort reason (or NULL) */ static void abort_for_stonith_failure(enum pcmk__graph_next abort_action, const char *target, const xmlNode *reason) { /* If stonith repeatedly fails, we eventually give up on starting a new * transition for that reason. */ if ((abort_action != pcmk__graph_wait) && too_many_st_failures(target)) { abort_action = pcmk__graph_wait; } abort_transition(INFINITY, abort_action, "Stonith failed", reason); } /* * stonith cleanup list * * If the DC is shot, proper notifications might not go out. * The stonith cleanup list allows the cluster to (re-)send * notifications once a new DC is elected. */ static GList *stonith_cleanup_list = NULL; /*! * \internal * \brief Add a node to the stonith cleanup list * * \param[in] target Name of node to add */ void add_stonith_cleanup(const char *target) { stonith_cleanup_list = g_list_append(stonith_cleanup_list, strdup(target)); } /*! * \internal * \brief Remove a node from the stonith cleanup list * * \param[in] Name of node to remove */ void remove_stonith_cleanup(const char *target) { GList *iter = stonith_cleanup_list; while (iter != NULL) { GList *tmp = iter; char *iter_name = tmp->data; iter = iter->next; if (pcmk__str_eq(target, iter_name, pcmk__str_casei)) { crm_trace("Removing %s from the cleanup list", iter_name); stonith_cleanup_list = g_list_delete_link(stonith_cleanup_list, tmp); free(iter_name); } } } /*! * \internal * \brief Purge all entries from the stonith cleanup list */ void purge_stonith_cleanup(void) { if (stonith_cleanup_list) { GList *iter = NULL; for (iter = stonith_cleanup_list; iter != NULL; iter = iter->next) { char *target = iter->data; crm_info("Purging %s from stonith cleanup list", target); free(target); } g_list_free(stonith_cleanup_list); stonith_cleanup_list = NULL; } } /*! * \internal * \brief Send stonith updates for all entries in cleanup list, then purge it */ void execute_stonith_cleanup(void) { GList *iter; for (iter = stonith_cleanup_list; iter != NULL; iter = iter->next) { char *target = iter->data; crm_node_t *target_node = crm_get_peer(0, target); const char *uuid = crm_peer_uuid(target_node); crm_notice("Marking %s, target of a previous stonith action, as clean", target); send_stonith_update(NULL, target, uuid); free(target); } g_list_free(stonith_cleanup_list); stonith_cleanup_list = NULL; } /* end stonith cleanup list functions */ /* stonith API client * * Functions that need to interact directly with the fencer via its API */ static stonith_t *stonith_api = NULL; static crm_trigger_t *stonith_reconnect = NULL; static char *te_client_id = NULL; static gboolean fail_incompletable_stonith(pcmk__graph_t *graph) { GList *lpc = NULL; const char *task = NULL; xmlNode *last_action = NULL; if (graph == NULL) { return FALSE; } for (lpc = graph->synapses; lpc != NULL; lpc = lpc->next) { GList *lpc2 = NULL; pcmk__graph_synapse_t *synapse = (pcmk__graph_synapse_t *) lpc->data; if (pcmk_is_set(synapse->flags, pcmk__synapse_confirmed)) { continue; } for (lpc2 = synapse->actions; lpc2 != NULL; lpc2 = lpc2->next) { pcmk__graph_action_t *action = (pcmk__graph_action_t *) lpc2->data; if ((action->type != pcmk__cluster_graph_action) || pcmk_is_set(action->flags, pcmk__graph_action_confirmed)) { continue; } task = crm_element_value(action->xml, XML_LRM_ATTR_TASK); if (task && pcmk__str_eq(task, CRM_OP_FENCE, pcmk__str_casei)) { pcmk__set_graph_action_flags(action, pcmk__graph_action_failed); last_action = action->xml; pcmk__update_graph(graph, action); crm_notice("Failing action %d (%s): fencer terminated", action->id, ID(action->xml)); } } } if (last_action != NULL) { crm_warn("Fencer failure resulted in unrunnable actions"); abort_for_stonith_failure(pcmk__graph_restart, NULL, last_action); return TRUE; } return FALSE; } static void tengine_stonith_connection_destroy(stonith_t *st, stonith_event_t *e) { te_cleanup_stonith_history_sync(st, FALSE); if (pcmk_is_set(controld_globals.fsa_input_register, R_ST_REQUIRED)) { crm_crit("Fencing daemon connection failed"); mainloop_set_trigger(stonith_reconnect); } else { crm_info("Fencing daemon disconnected"); } if (stonith_api) { /* the client API won't properly reconnect notifications * if they are still in the table - so remove them */ if (stonith_api->state != stonith_disconnected) { stonith_api->cmds->disconnect(st); } stonith_api->cmds->remove_notification(stonith_api, NULL); } if (AM_I_DC) { fail_incompletable_stonith(controld_globals.transition_graph); trigger_graph(); } } /*! * \internal * \brief Handle an event notification from the fencing API * * \param[in] st Fencing API connection (ignored) * \param[in] event Fencing API event notification */ static void handle_fence_notification(stonith_t *st, stonith_event_t *event) { bool succeeded = true; const char *executioner = "the cluster"; const char *client = "a client"; const char *reason = NULL; int exec_status; if (te_client_id == NULL) { te_client_id = crm_strdup_printf("%s.%lu", crm_system_name, (unsigned long) getpid()); } if (event == NULL) { crm_err("Notify data not found"); return; } if (event->executioner != NULL) { executioner = event->executioner; } if (event->client_origin != NULL) { client = event->client_origin; } exec_status = stonith__event_execution_status(event); if ((stonith__event_exit_status(event) != CRM_EX_OK) || (exec_status != PCMK_EXEC_DONE)) { succeeded = false; if (exec_status == PCMK_EXEC_DONE) { exec_status = PCMK_EXEC_ERROR; } } reason = stonith__event_exit_reason(event); crmd_alert_fencing_op(event); if (pcmk__str_eq("on", event->action, pcmk__str_none)) { // Unfencing doesn't need special handling, just a log message if (succeeded) { crm_notice("%s was unfenced by %s at the request of %s@%s", event->target, executioner, client, event->origin); } else { crm_err("Unfencing of %s by %s failed (%s%s%s) with exit status %d", event->target, executioner, pcmk_exec_status_str(exec_status), ((reason == NULL)? "" : ": "), ((reason == NULL)? "" : reason), stonith__event_exit_status(event)); } return; } if (succeeded && pcmk__str_eq(event->target, controld_globals.our_nodename, pcmk__str_casei)) { /* We were notified of our own fencing. Most likely, either fencing was * misconfigured, or fabric fencing that doesn't cut cluster * communication is in use. * * Either way, shutting down the local host is a good idea, to require * administrator intervention. Also, other nodes would otherwise likely * set our status to lost because of the fencing callback and discard * our subsequent election votes as "not part of our cluster". */ crm_crit("We were allegedly just fenced by %s for %s!", executioner, event->origin); // Dumps blackbox if enabled if (fence_reaction_panic) { pcmk__panic(__func__); } else { crm_exit(CRM_EX_FATAL); } return; // Should never get here } /* Update the count of fencing failures for this target, in case we become * DC later. The current DC has already updated its fail count in * tengine_stonith_callback(). */ if (!AM_I_DC) { if (succeeded) { st_fail_count_reset(event->target); } else { st_fail_count_increment(event->target); } } crm_notice("Peer %s was%s terminated (%s) by %s on behalf of %s@%s: " "%s%s%s%s " CRM_XS " event=%s", event->target, (succeeded? "" : " not"), event->action, executioner, client, event->origin, (succeeded? "OK" : pcmk_exec_status_str(exec_status)), ((reason == NULL)? "" : " ("), ((reason == NULL)? "" : reason), ((reason == NULL)? "" : ")"), event->id); if (succeeded) { crm_node_t *peer = pcmk__search_known_node_cache(0, event->target, CRM_GET_PEER_ANY); const char *uuid = NULL; if (peer == NULL) { return; } uuid = crm_peer_uuid(peer); if (AM_I_DC) { /* The DC always sends updates */ send_stonith_update(NULL, event->target, uuid); /* @TODO Ideally, at this point, we'd check whether the fenced node * hosted any guest nodes, and call remote_node_down() for them. * Unfortunately, the controller doesn't have a simple, reliable way * to map hosts to guests. It might be possible to track this in the * peer cache via crm_remote_peer_cache_refresh(). For now, we rely * on the scheduler creating fence pseudo-events for the guests. */ if (!pcmk__str_eq(client, te_client_id, pcmk__str_casei)) { /* Abort the current transition if it wasn't the cluster that * initiated fencing. */ crm_info("External fencing operation from %s fenced %s", client, event->target); abort_transition(INFINITY, pcmk__graph_restart, "External Fencing Operation", NULL); } } else if (pcmk__str_eq(controld_globals.dc_name, event->target, pcmk__str_null_matches|pcmk__str_casei) && !pcmk_is_set(peer->flags, crm_remote_node)) { // Assume the target was our DC if we don't currently have one if (controld_globals.dc_name != NULL) { crm_notice("Fencing target %s was our DC", event->target); } else { crm_notice("Fencing target %s may have been our DC", event->target); } /* Given the CIB resyncing that occurs around elections, * have one node update the CIB now and, if the new DC is different, * have them do so too after the election */ if (pcmk__str_eq(event->executioner, controld_globals.our_nodename, pcmk__str_casei)) { send_stonith_update(NULL, event->target, uuid); } add_stonith_cleanup(event->target); } /* If the target is a remote node, and we host its connection, * immediately fail all monitors so it can be recovered quickly. * The connection won't necessarily drop when a remote node is fenced, * so the failure might not otherwise be detected until the next poke. */ if (pcmk_is_set(peer->flags, crm_remote_node)) { remote_ra_fail(event->target); } crmd_peer_down(peer, TRUE); } } /*! * \brief Connect to fencer * * \param[in] user_data If NULL, retry failures now, otherwise retry in main loop * * \return TRUE * \note If user_data is NULL, this will wait 2s between attempts, for up to * 30 attempts, meaning the controller could be blocked as long as 58s. */ static gboolean te_connect_stonith(gpointer user_data) { int rc = pcmk_ok; if (stonith_api == NULL) { stonith_api = stonith_api_new(); if (stonith_api == NULL) { crm_err("Could not connect to fencer: API memory allocation failed"); return TRUE; } } if (stonith_api->state != stonith_disconnected) { crm_trace("Already connected to fencer, no need to retry"); return TRUE; } if (user_data == NULL) { // Blocking (retry failures now until successful) rc = stonith_api_connect_retry(stonith_api, crm_system_name, 30); if (rc != pcmk_ok) { crm_err("Could not connect to fencer in 30 attempts: %s " CRM_XS " rc=%d", pcmk_strerror(rc), rc); } } else { // Non-blocking (retry failures later in main loop) rc = stonith_api->cmds->connect(stonith_api, crm_system_name, NULL); if (rc != pcmk_ok) { if (pcmk_is_set(controld_globals.fsa_input_register, R_ST_REQUIRED)) { crm_notice("Fencer connection failed (will retry): %s " CRM_XS " rc=%d", pcmk_strerror(rc), rc); mainloop_set_trigger(stonith_reconnect); } else { crm_info("Fencer connection failed (ignoring because no longer required): %s " CRM_XS " rc=%d", pcmk_strerror(rc), rc); } return TRUE; } } if (rc == pcmk_ok) { stonith_api->cmds->register_notification(stonith_api, T_STONITH_NOTIFY_DISCONNECT, tengine_stonith_connection_destroy); stonith_api->cmds->register_notification(stonith_api, T_STONITH_NOTIFY_FENCE, handle_fence_notification); stonith_api->cmds->register_notification(stonith_api, T_STONITH_NOTIFY_HISTORY_SYNCED, tengine_stonith_history_synced); te_trigger_stonith_history_sync(TRUE); crm_notice("Fencer successfully connected"); } return TRUE; } /*! \internal \brief Schedule fencer connection attempt in main loop */ void controld_trigger_fencer_connect(void) { if (stonith_reconnect == NULL) { stonith_reconnect = mainloop_add_trigger(G_PRIORITY_LOW, te_connect_stonith, GINT_TO_POINTER(TRUE)); } controld_set_fsa_input_flags(R_ST_REQUIRED); mainloop_set_trigger(stonith_reconnect); } void controld_disconnect_fencer(bool destroy) { if (stonith_api) { // Prevent fencer connection from coming up again controld_clear_fsa_input_flags(R_ST_REQUIRED); if (stonith_api->state != stonith_disconnected) { stonith_api->cmds->disconnect(stonith_api); } stonith_api->cmds->remove_notification(stonith_api, NULL); } if (destroy) { if (stonith_api) { stonith_api->cmds->free(stonith_api); stonith_api = NULL; } if (stonith_reconnect) { mainloop_destroy_trigger(stonith_reconnect); stonith_reconnect = NULL; } if (te_client_id) { free(te_client_id); te_client_id = NULL; } } } static gboolean do_stonith_history_sync(gpointer user_data) { if (stonith_api && (stonith_api->state != stonith_disconnected)) { stonith_history_t *history = NULL; te_cleanup_stonith_history_sync(stonith_api, FALSE); stonith_api->cmds->history(stonith_api, st_opt_sync_call | st_opt_broadcast, NULL, &history, 5); stonith_history_free(history); return TRUE; } else { crm_info("Skip triggering stonith history-sync as stonith is disconnected"); return FALSE; } } static void tengine_stonith_callback(stonith_t *stonith, stonith_callback_data_t *data) { char *uuid = NULL; int stonith_id = -1; int transition_id = -1; pcmk__graph_action_t *action = NULL; const char *target = NULL; if ((data == NULL) || (data->userdata == NULL)) { crm_err("Ignoring fence operation %d result: " "No transition key given (bug?)", ((data == NULL)? -1 : data->call_id)); return; } if (!AM_I_DC) { const char *reason = stonith__exit_reason(data); if (reason == NULL) { reason = pcmk_exec_status_str(stonith__execution_status(data)); } crm_notice("Result of fence operation %d: %d (%s) " CRM_XS " key=%s", data->call_id, stonith__exit_status(data), reason, (const char *) data->userdata); return; } CRM_CHECK(decode_transition_key(data->userdata, &uuid, &transition_id, &stonith_id, NULL), goto bail); if (controld_globals.transition_graph->complete || (stonith_id < 0) || !pcmk__str_eq(uuid, controld_globals.te_uuid, pcmk__str_none) || (controld_globals.transition_graph->id != transition_id)) { crm_info("Ignoring fence operation %d result: " "Not from current transition " CRM_XS " complete=%s action=%d uuid=%s (vs %s) transition=%d (vs %d)", data->call_id, pcmk__btoa(controld_globals.transition_graph->complete), stonith_id, uuid, controld_globals.te_uuid, transition_id, controld_globals.transition_graph->id); goto bail; } action = controld_get_action(stonith_id); if (action == NULL) { crm_err("Ignoring fence operation %d result: " "Action %d not found in transition graph (bug?) " CRM_XS " uuid=%s transition=%d", data->call_id, stonith_id, uuid, transition_id); goto bail; } target = crm_element_value(action->xml, XML_LRM_ATTR_TARGET); if (target == NULL) { crm_err("Ignoring fence operation %d result: No target given (bug?)", data->call_id); goto bail; } stop_te_timer(action); if (stonith__exit_status(data) == CRM_EX_OK) { const char *uuid = crm_element_value(action->xml, XML_LRM_ATTR_TARGET_UUID); const char *op = crm_meta_value(action->params, "stonith_action"); crm_info("Fence operation %d for %s succeeded", data->call_id, target); if (!(pcmk_is_set(action->flags, pcmk__graph_action_confirmed))) { te_action_confirmed(action, NULL); if (pcmk__str_eq("on", op, pcmk__str_casei)) { const char *value = NULL; char *now = pcmk__ttoa(time(NULL)); gboolean is_remote_node = FALSE; /* This check is not 100% reliable, since this node is not * guaranteed to have the remote node cached. However, it * doesn't have to be reliable, since the attribute manager can * learn a node's "remoteness" by other means sooner or later. * This allows it to learn more quickly if this node does have * the information. */ if (g_hash_table_lookup(crm_remote_peer_cache, uuid) != NULL) { is_remote_node = TRUE; } update_attrd(target, CRM_ATTR_UNFENCED, now, NULL, is_remote_node); free(now); value = crm_meta_value(action->params, XML_OP_ATTR_DIGESTS_ALL); update_attrd(target, CRM_ATTR_DIGESTS_ALL, value, NULL, is_remote_node); value = crm_meta_value(action->params, XML_OP_ATTR_DIGESTS_SECURE); update_attrd(target, CRM_ATTR_DIGESTS_SECURE, value, NULL, is_remote_node); } else if (!(pcmk_is_set(action->flags, pcmk__graph_action_sent_update))) { send_stonith_update(action, target, uuid); pcmk__set_graph_action_flags(action, pcmk__graph_action_sent_update); } } st_fail_count_reset(target); } else { enum pcmk__graph_next abort_action = pcmk__graph_restart; int status = stonith__execution_status(data); const char *reason = stonith__exit_reason(data); if (reason == NULL) { if (status == PCMK_EXEC_DONE) { reason = "Agent returned error"; } else { reason = pcmk_exec_status_str(status); } } pcmk__set_graph_action_flags(action, pcmk__graph_action_failed); /* If no fence devices were available, there's no use in immediately * checking again, so don't start a new transition in that case. */ if (status == PCMK_EXEC_NO_FENCE_DEVICE) { crm_warn("Fence operation %d for %s failed: %s " "(aborting transition and giving up for now)", data->call_id, target, reason); abort_action = pcmk__graph_wait; } else { crm_notice("Fence operation %d for %s failed: %s " "(aborting transition)", data->call_id, target, reason); } /* Increment the fail count now, so abort_for_stonith_failure() can * check it. Non-DC nodes will increment it in * handle_fence_notification(). */ st_fail_count_increment(target); abort_for_stonith_failure(abort_action, target, NULL); } pcmk__update_graph(controld_globals.transition_graph, action); trigger_graph(); bail: free(data->userdata); free(uuid); return; } static int fence_with_delay(const char *target, const char *type, const char *delay) { uint32_t options = st_opt_none; // Group of enum stonith_call_options int timeout_sec = (int) (controld_globals.transition_graph->stonith_timeout / 1000); int delay_i; if (crmd_join_phase_count(crm_join_confirmed) == 1) { stonith__set_call_options(options, target, st_opt_allow_suicide); } pcmk__scan_min_int(delay, &delay_i, 0); return stonith_api->cmds->fence_with_delay(stonith_api, options, target, type, timeout_sec, 0, delay_i); } /*! * \internal * \brief Execute a fencing action from a transition graph * * \param[in] graph Transition graph being executed (ignored) * \param[in] action Fencing action to execute * * \return Standard Pacemaker return code */ int controld_execute_fence_action(pcmk__graph_t *graph, pcmk__graph_action_t *action) { int rc = 0; const char *id = ID(action->xml); const char *uuid = crm_element_value(action->xml, XML_LRM_ATTR_TARGET_UUID); const char *target = crm_element_value(action->xml, XML_LRM_ATTR_TARGET); const char *type = crm_meta_value(action->params, "stonith_action"); char *transition_key = NULL; const char *priority_delay = NULL; gboolean invalid_action = FALSE; guint stonith_timeout = controld_globals.transition_graph->stonith_timeout; CRM_CHECK(id != NULL, invalid_action = TRUE); CRM_CHECK(uuid != NULL, invalid_action = TRUE); CRM_CHECK(type != NULL, invalid_action = TRUE); CRM_CHECK(target != NULL, invalid_action = TRUE); if (invalid_action) { crm_log_xml_warn(action->xml, "BadAction"); return EPROTO; } priority_delay = crm_meta_value(action->params, XML_CONFIG_ATTR_PRIORITY_FENCING_DELAY); crm_notice("Requesting fencing (%s) of node %s " CRM_XS " action=%s timeout=%u%s%s", type, target, id, stonith_timeout, priority_delay ? " priority_delay=" : "", priority_delay ? priority_delay : ""); /* Passing NULL means block until we can connect... */ te_connect_stonith(NULL); rc = fence_with_delay(target, type, priority_delay); transition_key = pcmk__transition_key(controld_globals.transition_graph->id, action->id, 0, controld_globals.te_uuid), stonith_api->cmds->register_callback(stonith_api, rc, (int) (stonith_timeout / 1000), st_opt_timeout_updates, transition_key, "tengine_stonith_callback", tengine_stonith_callback); return pcmk_rc_ok; } bool controld_verify_stonith_watchdog_timeout(const char *value) { const char *our_nodename = controld_globals.our_nodename; gboolean rv = TRUE; if (stonith_api && (stonith_api->state != stonith_disconnected) && stonith__watchdog_fencing_enabled_for_node_api(stonith_api, our_nodename)) { rv = pcmk__valid_sbd_timeout(value); } return rv; } /* end stonith API client functions */ /* * stonith history synchronization * * Each node's fencer keeps track of a cluster-wide fencing history. When a node * joins or leaves, we need to synchronize the history across all nodes. */ static crm_trigger_t *stonith_history_sync_trigger = NULL; static mainloop_timer_t *stonith_history_sync_timer_short = NULL; static mainloop_timer_t *stonith_history_sync_timer_long = NULL; void te_cleanup_stonith_history_sync(stonith_t *st, bool free_timers) { if (free_timers) { mainloop_timer_del(stonith_history_sync_timer_short); stonith_history_sync_timer_short = NULL; mainloop_timer_del(stonith_history_sync_timer_long); stonith_history_sync_timer_long = NULL; } else { mainloop_timer_stop(stonith_history_sync_timer_short); mainloop_timer_stop(stonith_history_sync_timer_long); } if (st) { st->cmds->remove_notification(st, T_STONITH_NOTIFY_HISTORY_SYNCED); } } static void tengine_stonith_history_synced(stonith_t *st, stonith_event_t *st_event) { te_cleanup_stonith_history_sync(st, FALSE); crm_debug("Fence-history synced - cancel all timers"); } static gboolean stonith_history_sync_set_trigger(gpointer user_data) { mainloop_set_trigger(stonith_history_sync_trigger); return FALSE; } void te_trigger_stonith_history_sync(bool long_timeout) { /* trigger a sync in 5s to give more nodes the * chance to show up so that we don't create * unnecessary stonith-history-sync traffic * * the long timeout of 30s is there as a fallback * so that after a successful connection to fenced * we will wait for 30s for the DC to trigger a * history-sync * if this doesn't happen we trigger a sync locally * (e.g. fenced segfaults and is restarted by pacemakerd) */ /* as we are finally checking the stonith-connection * in do_stonith_history_sync we should be fine * leaving stonith_history_sync_time & stonith_history_sync_trigger * around */ if (stonith_history_sync_trigger == NULL) { stonith_history_sync_trigger = mainloop_add_trigger(G_PRIORITY_LOW, do_stonith_history_sync, NULL); } if (long_timeout) { if(stonith_history_sync_timer_long == NULL) { stonith_history_sync_timer_long = mainloop_timer_add("history_sync_long", 30000, FALSE, stonith_history_sync_set_trigger, NULL); } crm_info("Fence history will be synchronized cluster-wide within 30 seconds"); mainloop_timer_start(stonith_history_sync_timer_long); } else { if(stonith_history_sync_timer_short == NULL) { stonith_history_sync_timer_short = mainloop_timer_add("history_sync_short", 5000, FALSE, stonith_history_sync_set_trigger, NULL); } crm_info("Fence history will be synchronized cluster-wide within 5 seconds"); mainloop_timer_start(stonith_history_sync_timer_short); } } /* end stonith history synchronization functions */ diff --git a/daemons/controld/controld_te_actions.c b/daemons/controld/controld_te_actions.c index a5629223ec..d8cfcad489 100644 --- a/daemons/controld/controld_te_actions.c +++ b/daemons/controld/controld_te_actions.c @@ -1,746 +1,746 @@ /* * Copyright 2004-2023 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 // lrmd_event_data_t, lrmd_free_event() #include #include #include #include #include static GHashTable *te_targets = NULL; void send_rsc_command(pcmk__graph_action_t *action); static void te_update_job_count(pcmk__graph_action_t *action, int offset); static void te_start_action_timer(const pcmk__graph_t *graph, pcmk__graph_action_t *action) { action->timer = g_timeout_add(action->timeout + graph->network_delay, action_timer_callback, (void *) action); CRM_ASSERT(action->timer != 0); } /*! * \internal * \brief Execute a graph pseudo-action * * \param[in,out] graph Transition graph being executed * \param[in,out] pseudo Pseudo-action to execute * * \return Standard Pacemaker return code */ static int execute_pseudo_action(pcmk__graph_t *graph, pcmk__graph_action_t *pseudo) { const char *task = crm_element_value(pseudo->xml, XML_LRM_ATTR_TASK); /* send to peers as well? */ if (pcmk__str_eq(task, CRM_OP_MAINTENANCE_NODES, pcmk__str_casei)) { GHashTableIter iter; crm_node_t *node = NULL; g_hash_table_iter_init(&iter, crm_peer_cache); while (g_hash_table_iter_next(&iter, NULL, (gpointer *) &node)) { xmlNode *cmd = NULL; if (pcmk__str_eq(controld_globals.our_nodename, node->uname, pcmk__str_casei)) { continue; } cmd = create_request(task, pseudo->xml, node->uname, CRM_SYSTEM_CRMD, CRM_SYSTEM_TENGINE, NULL); send_cluster_message(node, crm_msg_crmd, cmd, FALSE); free_xml(cmd); } remote_ra_process_maintenance_nodes(pseudo->xml); } else { /* Check action for Pacemaker Remote node side effects */ remote_ra_process_pseudo(pseudo->xml); } crm_debug("Pseudo-action %d (%s) fired and confirmed", pseudo->id, crm_element_value(pseudo->xml, XML_LRM_ATTR_TASK_KEY)); te_action_confirmed(pseudo, graph); return pcmk_rc_ok; } static int get_target_rc(pcmk__graph_action_t *action) { int exit_status; pcmk__scan_min_int(crm_meta_value(action->params, XML_ATTR_TE_TARGET_RC), &exit_status, 0); return exit_status; } /*! * \internal * \brief Execute a cluster action from a transition graph * * \param[in,out] graph Transition graph being executed * \param[in,out] action Cluster action to execute * * \return Standard Pacemaker return code */ static int execute_cluster_action(pcmk__graph_t *graph, pcmk__graph_action_t *action) { char *counter = NULL; xmlNode *cmd = NULL; gboolean is_local = FALSE; const char *id = NULL; const char *task = NULL; const char *value = NULL; const char *on_node = NULL; const char *router_node = NULL; gboolean rc = TRUE; gboolean no_wait = FALSE; id = ID(action->xml); CRM_CHECK(!pcmk__str_empty(id), return EPROTO); task = crm_element_value(action->xml, XML_LRM_ATTR_TASK); CRM_CHECK(!pcmk__str_empty(task), return EPROTO); on_node = crm_element_value(action->xml, XML_LRM_ATTR_TARGET); CRM_CHECK(!pcmk__str_empty(on_node), return pcmk_rc_node_unknown); router_node = crm_element_value(action->xml, XML_LRM_ATTR_ROUTER_NODE); if (router_node == NULL) { router_node = on_node; if (pcmk__str_eq(task, CRM_OP_LRM_DELETE, pcmk__str_none)) { const char *mode = crm_element_value(action->xml, PCMK__XA_MODE); if (pcmk__str_eq(mode, XML_TAG_CIB, pcmk__str_none)) { router_node = controld_globals.our_nodename; } } } if (pcmk__str_eq(router_node, controld_globals.our_nodename, pcmk__str_casei)) { is_local = TRUE; } value = crm_meta_value(action->params, XML_ATTR_TE_NOWAIT); if (crm_is_true(value)) { no_wait = TRUE; } crm_info("Handling controller request '%s' (%s on %s)%s%s", id, task, on_node, (is_local? " locally" : ""), (no_wait? " without waiting" : "")); if (is_local && pcmk__str_eq(task, CRM_OP_SHUTDOWN, pcmk__str_none)) { /* defer until everything else completes */ crm_info("Controller request '%s' is a local shutdown", id); graph->completion_action = pcmk__graph_shutdown; graph->abort_reason = "local shutdown"; te_action_confirmed(action, graph); return pcmk_rc_ok; } else if (pcmk__str_eq(task, CRM_OP_SHUTDOWN, pcmk__str_none)) { crm_node_t *peer = crm_get_peer(0, router_node); pcmk__update_peer_expected(__func__, peer, CRMD_JOINSTATE_DOWN); } cmd = create_request(task, action->xml, router_node, CRM_SYSTEM_CRMD, CRM_SYSTEM_TENGINE, NULL); counter = pcmk__transition_key(controld_globals.transition_graph->id, action->id, get_target_rc(action), controld_globals.te_uuid); crm_xml_add(cmd, XML_ATTR_TRANSITION_KEY, counter); rc = send_cluster_message(crm_get_peer(0, router_node), crm_msg_crmd, cmd, TRUE); free(counter); free_xml(cmd); if (rc == FALSE) { crm_err("Action %d failed: send", action->id); return ECOMM; } else if (no_wait) { te_action_confirmed(action, graph); } else { if (action->timeout <= 0) { crm_err("Action %d: %s on %s had an invalid timeout (%dms). Using %ums instead", action->id, task, on_node, action->timeout, graph->network_delay); action->timeout = (int) graph->network_delay; } te_start_action_timer(graph, action); } return pcmk_rc_ok; } /*! * \internal * \brief Synthesize an executor event for a resource action timeout * * \param[in] action Resource action that timed out * \param[in] target_rc Expected result of action that timed out * * Synthesize an executor event for a resource action timeout. (If the executor * gets a timeout while waiting for a resource action to complete, that will be * reported via the usual callback. This timeout means we didn't hear from the * executor itself or the controller that relayed the action to the executor.) * * \return Newly created executor event for result of \p action * \note The caller is responsible for freeing the return value using * lrmd_free_event(). */ static lrmd_event_data_t * synthesize_timeout_event(const pcmk__graph_action_t *action, int target_rc) { lrmd_event_data_t *op = NULL; const char *target = crm_element_value(action->xml, XML_LRM_ATTR_TARGET); const char *reason = NULL; char *dynamic_reason = NULL; if (pcmk__str_eq(target, get_local_node_name(), pcmk__str_casei)) { reason = "Local executor did not return result in time"; } else { const char *router_node = NULL; router_node = crm_element_value(action->xml, XML_LRM_ATTR_ROUTER_NODE); if (router_node == NULL) { router_node = target; } dynamic_reason = crm_strdup_printf("Controller on %s did not return " "result in time", router_node); reason = dynamic_reason; } op = pcmk__event_from_graph_action(NULL, action, PCMK_EXEC_TIMEOUT, PCMK_OCF_UNKNOWN_ERROR, reason); op->call_id = -1; op->user_data = pcmk__transition_key(controld_globals.transition_graph->id, action->id, target_rc, controld_globals.te_uuid); free(dynamic_reason); return op; } static void controld_record_action_event(pcmk__graph_action_t *action, lrmd_event_data_t *op) { cib_t *cib_conn = controld_globals.cib_conn; xmlNode *state = NULL; xmlNode *rsc = NULL; xmlNode *action_rsc = NULL; int rc = pcmk_ok; const char *rsc_id = NULL; const char *target = crm_element_value(action->xml, XML_LRM_ATTR_TARGET); const char *task_uuid = crm_element_value(action->xml, XML_LRM_ATTR_TASK_KEY); const char *target_uuid = crm_element_value(action->xml, XML_LRM_ATTR_TARGET_UUID); int target_rc = get_target_rc(action); action_rsc = find_xml_node(action->xml, XML_CIB_TAG_RESOURCE, TRUE); if (action_rsc == NULL) { return; } rsc_id = ID(action_rsc); CRM_CHECK(rsc_id != NULL, crm_log_xml_err(action->xml, "Bad:action"); return); /* update the CIB */ state = create_xml_node(NULL, XML_CIB_TAG_STATE); crm_xml_add(state, XML_ATTR_ID, target_uuid); crm_xml_add(state, XML_ATTR_UNAME, target); rsc = create_xml_node(state, XML_CIB_TAG_LRM); crm_xml_add(rsc, XML_ATTR_ID, target_uuid); rsc = create_xml_node(rsc, XML_LRM_TAG_RESOURCES); rsc = create_xml_node(rsc, XML_LRM_TAG_RESOURCE); crm_xml_add(rsc, XML_ATTR_ID, rsc_id); crm_copy_xml_element(action_rsc, rsc, XML_ATTR_TYPE); crm_copy_xml_element(action_rsc, rsc, XML_AGENT_ATTR_CLASS); crm_copy_xml_element(action_rsc, rsc, XML_AGENT_ATTR_PROVIDER); pcmk__create_history_xml(rsc, op, CRM_FEATURE_SET, target_rc, target, __func__); - rc = cib_conn->cmds->update(cib_conn, XML_CIB_TAG_STATUS, state, + rc = cib_conn->cmds->modify(cib_conn, XML_CIB_TAG_STATUS, state, cib_scope_local); fsa_register_cib_callback(rc, NULL, cib_action_updated); free_xml(state); crm_trace("Sent CIB update (call ID %d) for synthesized event of action %d (%s on %s)", rc, action->id, task_uuid, target); pcmk__set_graph_action_flags(action, pcmk__graph_action_sent_update); } void controld_record_action_timeout(pcmk__graph_action_t *action) { lrmd_event_data_t *op = NULL; const char *target = crm_element_value(action->xml, XML_LRM_ATTR_TARGET); const char *task_uuid = crm_element_value(action->xml, XML_LRM_ATTR_TASK_KEY); int target_rc = get_target_rc(action); crm_warn("%s %d: %s on %s timed out", crm_element_name(action->xml), action->id, task_uuid, target); op = synthesize_timeout_event(action, target_rc); controld_record_action_event(action, op); lrmd_free_event(op); } /*! * \internal * \brief Execute a resource action from a transition graph * * \param[in,out] graph Transition graph being executed * \param[in,out] action Resource action to execute * * \return Standard Pacemaker return code */ static int execute_rsc_action(pcmk__graph_t *graph, pcmk__graph_action_t *action) { /* never overwrite stop actions in the CIB with * anything other than completed results * * Writing pending stops makes it look like the * resource is running again */ xmlNode *cmd = NULL; xmlNode *rsc_op = NULL; gboolean rc = TRUE; gboolean no_wait = FALSE; gboolean is_local = FALSE; char *counter = NULL; const char *task = NULL; const char *value = NULL; const char *on_node = NULL; const char *router_node = NULL; const char *task_uuid = NULL; CRM_ASSERT(action != NULL); CRM_ASSERT(action->xml != NULL); pcmk__clear_graph_action_flags(action, pcmk__graph_action_executed); on_node = crm_element_value(action->xml, XML_LRM_ATTR_TARGET); CRM_CHECK(!pcmk__str_empty(on_node), crm_err("Corrupted command(id=%s) %s: no node", ID(action->xml), pcmk__s(task, "without task")); return pcmk_rc_node_unknown); rsc_op = action->xml; task = crm_element_value(rsc_op, XML_LRM_ATTR_TASK); task_uuid = crm_element_value(action->xml, XML_LRM_ATTR_TASK_KEY); router_node = crm_element_value(rsc_op, XML_LRM_ATTR_ROUTER_NODE); if (!router_node) { router_node = on_node; } counter = pcmk__transition_key(controld_globals.transition_graph->id, action->id, get_target_rc(action), controld_globals.te_uuid); crm_xml_add(rsc_op, XML_ATTR_TRANSITION_KEY, counter); if (pcmk__str_eq(router_node, controld_globals.our_nodename, pcmk__str_casei)) { is_local = TRUE; } value = crm_meta_value(action->params, XML_ATTR_TE_NOWAIT); if (crm_is_true(value)) { no_wait = TRUE; } crm_notice("Initiating %s operation %s%s on %s%s "CRM_XS" action %d", task, task_uuid, (is_local? " locally" : ""), on_node, (no_wait? " without waiting" : ""), action->id); cmd = create_request(CRM_OP_INVOKE_LRM, rsc_op, router_node, CRM_SYSTEM_LRMD, CRM_SYSTEM_TENGINE, NULL); if (is_local) { /* shortcut local resource commands */ ha_msg_input_t data = { .msg = cmd, .xml = rsc_op, }; fsa_data_t msg = { .id = 0, .data = &data, .data_type = fsa_dt_ha_msg, .fsa_input = I_NULL, .fsa_cause = C_FSA_INTERNAL, .actions = A_LRM_INVOKE, .origin = __func__, }; do_lrm_invoke(A_LRM_INVOKE, C_FSA_INTERNAL, controld_globals.fsa_state, I_NULL, &msg); } else { rc = send_cluster_message(crm_get_peer(0, router_node), crm_msg_lrmd, cmd, TRUE); } free(counter); free_xml(cmd); pcmk__set_graph_action_flags(action, pcmk__graph_action_executed); if (rc == FALSE) { crm_err("Action %d failed: send", action->id); return ECOMM; } else if (no_wait) { /* Just mark confirmed. Don't bump the job count only to immediately * decrement it. */ crm_info("Action %d confirmed - no wait", action->id); pcmk__set_graph_action_flags(action, pcmk__graph_action_confirmed); pcmk__update_graph(controld_globals.transition_graph, action); trigger_graph(); } else if (pcmk_is_set(action->flags, pcmk__graph_action_confirmed)) { crm_debug("Action %d: %s %s on %s(timeout %dms) was already confirmed.", action->id, task, task_uuid, on_node, action->timeout); } else { if (action->timeout <= 0) { crm_err("Action %d: %s %s on %s had an invalid timeout (%dms). Using %ums instead", action->id, task, task_uuid, on_node, action->timeout, graph->network_delay); action->timeout = (int) graph->network_delay; } te_update_job_count(action, 1); te_start_action_timer(graph, action); } return pcmk_rc_ok; } struct te_peer_s { char *name; int jobs; int migrate_jobs; }; static void te_peer_free(gpointer p) { struct te_peer_s *peer = p; free(peer->name); free(peer); } void te_reset_job_counts(void) { GHashTableIter iter; struct te_peer_s *peer = NULL; if(te_targets == NULL) { te_targets = pcmk__strkey_table(NULL, te_peer_free); } g_hash_table_iter_init(&iter, te_targets); while (g_hash_table_iter_next(&iter, NULL, (gpointer *) & peer)) { peer->jobs = 0; peer->migrate_jobs = 0; } } static void te_update_job_count_on(const char *target, int offset, bool migrate) { struct te_peer_s *r = NULL; if(target == NULL || te_targets == NULL) { return; } r = g_hash_table_lookup(te_targets, target); if(r == NULL) { r = calloc(1, sizeof(struct te_peer_s)); r->name = strdup(target); g_hash_table_insert(te_targets, r->name, r); } r->jobs += offset; if(migrate) { r->migrate_jobs += offset; } crm_trace("jobs[%s] = %d", target, r->jobs); } static void te_update_job_count(pcmk__graph_action_t *action, int offset) { const char *task = crm_element_value(action->xml, XML_LRM_ATTR_TASK); const char *target = crm_element_value(action->xml, XML_LRM_ATTR_TARGET); if ((action->type != pcmk__rsc_graph_action) || (target == NULL)) { /* No limit on these */ return; } /* if we have a router node, this means the action is performing * on a remote node. For now, we count all actions occurring on a * remote node against the job list on the cluster node hosting * the connection resources */ target = crm_element_value(action->xml, XML_LRM_ATTR_ROUTER_NODE); if ((target == NULL) && pcmk__strcase_any_of(task, CRMD_ACTION_MIGRATE, CRMD_ACTION_MIGRATED, NULL)) { const char *t1 = crm_meta_value(action->params, XML_LRM_ATTR_MIGRATE_SOURCE); const char *t2 = crm_meta_value(action->params, XML_LRM_ATTR_MIGRATE_TARGET); te_update_job_count_on(t1, offset, TRUE); te_update_job_count_on(t2, offset, TRUE); return; } else if (target == NULL) { target = crm_element_value(action->xml, XML_LRM_ATTR_TARGET); } te_update_job_count_on(target, offset, FALSE); } /*! * \internal * \brief Check whether a graph action is allowed to be executed on a node * * \param[in] graph Transition graph being executed * \param[in] action Graph action being executed * \param[in] target Name of node where action should be executed * * \return true if action is allowed, otherwise false */ static bool allowed_on_node(const pcmk__graph_t *graph, const pcmk__graph_action_t *action, const char *target) { int limit = 0; struct te_peer_s *r = NULL; const char *task = crm_element_value(action->xml, XML_LRM_ATTR_TASK); const char *id = crm_element_value(action->xml, XML_LRM_ATTR_TASK_KEY); if(target == NULL) { /* No limit on these */ return true; } else if(te_targets == NULL) { return false; } r = g_hash_table_lookup(te_targets, target); limit = throttle_get_job_limit(target); if(r == NULL) { r = calloc(1, sizeof(struct te_peer_s)); r->name = strdup(target); g_hash_table_insert(te_targets, r->name, r); } if(limit <= r->jobs) { crm_trace("Peer %s is over their job limit of %d (%d): deferring %s", target, limit, r->jobs, id); return false; } else if(graph->migration_limit > 0 && r->migrate_jobs >= graph->migration_limit) { if (pcmk__strcase_any_of(task, CRMD_ACTION_MIGRATE, CRMD_ACTION_MIGRATED, NULL)) { crm_trace("Peer %s is over their migration job limit of %d (%d): deferring %s", target, graph->migration_limit, r->migrate_jobs, id); return false; } } crm_trace("Peer %s has not hit their limit yet. current jobs = %d limit= %d limit", target, r->jobs, limit); return true; } /*! * \internal * \brief Check whether a graph action is allowed to be executed * * \param[in] graph Transition graph being executed * \param[in] action Graph action being executed * * \return true if action is allowed, otherwise false */ static bool graph_action_allowed(pcmk__graph_t *graph, pcmk__graph_action_t *action) { const char *target = NULL; const char *task = crm_element_value(action->xml, XML_LRM_ATTR_TASK); if (action->type != pcmk__rsc_graph_action) { /* No limit on these */ return true; } /* if we have a router node, this means the action is performing * on a remote node. For now, we count all actions occurring on a * remote node against the job list on the cluster node hosting * the connection resources */ target = crm_element_value(action->xml, XML_LRM_ATTR_ROUTER_NODE); if ((target == NULL) && pcmk__strcase_any_of(task, CRMD_ACTION_MIGRATE, CRMD_ACTION_MIGRATED, NULL)) { target = crm_meta_value(action->params, XML_LRM_ATTR_MIGRATE_SOURCE); if (!allowed_on_node(graph, action, target)) { return false; } target = crm_meta_value(action->params, XML_LRM_ATTR_MIGRATE_TARGET); } else if (target == NULL) { target = crm_element_value(action->xml, XML_LRM_ATTR_TARGET); } return allowed_on_node(graph, action, target); } /*! * \brief Confirm a graph action (and optionally update graph) * * \param[in,out] action Action to confirm * \param[in,out] graph Update and trigger this graph (if non-NULL) */ void te_action_confirmed(pcmk__graph_action_t *action, pcmk__graph_t *graph) { if (!pcmk_is_set(action->flags, pcmk__graph_action_confirmed)) { if ((action->type == pcmk__rsc_graph_action) && (crm_element_value(action->xml, XML_LRM_ATTR_TARGET) != NULL)) { te_update_job_count(action, -1); } pcmk__set_graph_action_flags(action, pcmk__graph_action_confirmed); } if (graph) { pcmk__update_graph(graph, action); trigger_graph(); } } static pcmk__graph_functions_t te_graph_fns = { execute_pseudo_action, execute_rsc_action, execute_cluster_action, controld_execute_fence_action, graph_action_allowed, }; /* * \internal * \brief Register the transitioner's graph functions with \p libpacemaker */ void controld_register_graph_functions(void) { pcmk__set_graph_functions(&te_graph_fns); } void notify_crmd(pcmk__graph_t *graph) { const char *type = "unknown"; enum crmd_fsa_input event = I_NULL; crm_debug("Processing transition completion in state %s", fsa_state2string(controld_globals.fsa_state)); CRM_CHECK(graph->complete, graph->complete = true); switch (graph->completion_action) { case pcmk__graph_wait: type = "stop"; if (controld_globals.fsa_state == S_TRANSITION_ENGINE) { event = I_TE_SUCCESS; } break; case pcmk__graph_done: type = "done"; if (controld_globals.fsa_state == S_TRANSITION_ENGINE) { event = I_TE_SUCCESS; } break; case pcmk__graph_restart: type = "restart"; if (controld_globals.fsa_state == S_TRANSITION_ENGINE) { if (controld_get_period_transition_timer() > 0) { controld_stop_transition_timer(); controld_start_transition_timer(); } else { event = I_PE_CALC; } } else if (controld_globals.fsa_state == S_POLICY_ENGINE) { controld_set_fsa_action_flags(A_PE_INVOKE); controld_trigger_fsa(); } break; case pcmk__graph_shutdown: type = "shutdown"; if (pcmk_is_set(controld_globals.fsa_input_register, R_SHUTDOWN)) { event = I_STOP; } else { crm_err("We didn't ask to be shut down, yet the scheduler is telling us to"); event = I_TERMINATE; } } crm_debug("Transition %d status: %s - %s", graph->id, type, pcmk__s(graph->abort_reason, "unspecified reason")); graph->abort_reason = NULL; graph->completion_action = pcmk__graph_done; if (event != I_NULL) { register_fsa_input(C_FSA_INTERNAL, event, NULL); } else { controld_trigger_fsa(); } } diff --git a/lib/cib/cib_client.c b/lib/cib/cib_client.c index 70bcd7ed3d..adf731949b 100644 --- a/lib/cib/cib_client.c +++ b/lib/cib/cib_client.c @@ -1,758 +1,750 @@ /* * Copyright 2004-2022 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 static GHashTable *cib_op_callback_table = NULL; #define op_common(cib) do { \ if(cib == NULL) { \ return -EINVAL; \ } else if(cib->delegate_fn == NULL) { \ return -EPROTONOSUPPORT; \ } \ } while(0) static int cib_client_set_op_callback(cib_t *cib, void (*callback) (const xmlNode * msg, int call_id, int rc, xmlNode * output)) { if (callback == NULL) { crm_info("Un-Setting operation callback"); } else { crm_trace("Setting operation callback"); } cib->op_callback = callback; return pcmk_ok; } static gint ciblib_GCompareFunc(gconstpointer a, gconstpointer b) { int rc = 0; const cib_notify_client_t *a_client = a; const cib_notify_client_t *b_client = b; CRM_CHECK(a_client->event != NULL && b_client->event != NULL, return 0); rc = strcmp(a_client->event, b_client->event); if (rc == 0) { if (a_client->callback == b_client->callback) { return 0; } else if (((long)a_client->callback) < ((long)b_client->callback)) { crm_trace("callbacks for %s are not equal: %p < %p", a_client->event, a_client->callback, b_client->callback); return -1; } crm_trace("callbacks for %s are not equal: %p > %p", a_client->event, a_client->callback, b_client->callback); return 1; } return rc; } static int cib_client_add_notify_callback(cib_t * cib, const char *event, void (*callback) (const char *event, xmlNode * msg)) { GList *list_item = NULL; cib_notify_client_t *new_client = NULL; if ((cib->variant != cib_native) && (cib->variant != cib_remote)) { return -EPROTONOSUPPORT; } crm_trace("Adding callback for %s events (%d)", event, g_list_length(cib->notify_list)); new_client = calloc(1, sizeof(cib_notify_client_t)); new_client->event = event; new_client->callback = callback; list_item = g_list_find_custom(cib->notify_list, new_client, ciblib_GCompareFunc); if (list_item != NULL) { crm_warn("Callback already present"); free(new_client); return -EINVAL; } else { cib->notify_list = g_list_append(cib->notify_list, new_client); cib->cmds->register_notification(cib, event, 1); crm_trace("Callback added (%d)", g_list_length(cib->notify_list)); } return pcmk_ok; } static int get_notify_list_event_count(cib_t *cib, const char *event) { int count = 0; for (GList *iter = g_list_first(cib->notify_list); iter != NULL; iter = iter->next) { cib_notify_client_t *client = (cib_notify_client_t *) iter->data; if (strcmp(client->event, event) == 0) { count++; } } crm_trace("event(%s) count : %d", event, count); return count; } static int cib_client_del_notify_callback(cib_t *cib, const char *event, void (*callback) (const char *event, xmlNode *msg)) { GList *list_item = NULL; cib_notify_client_t *new_client = NULL; if (cib->variant != cib_native && cib->variant != cib_remote) { return -EPROTONOSUPPORT; } if (get_notify_list_event_count(cib, event) == 0) { crm_debug("The callback of the event does not exist(%s)", event); return pcmk_ok; } crm_debug("Removing callback for %s events", event); new_client = calloc(1, sizeof(cib_notify_client_t)); new_client->event = event; new_client->callback = callback; list_item = g_list_find_custom(cib->notify_list, new_client, ciblib_GCompareFunc); if (list_item != NULL) { cib_notify_client_t *list_client = list_item->data; cib->notify_list = g_list_remove(cib->notify_list, list_client); free(list_client); crm_trace("Removed callback"); } else { crm_trace("Callback not present"); } if (get_notify_list_event_count(cib, event) == 0) { /* When there is not the registration of the event, the processing turns off a notice. */ cib->cmds->register_notification(cib, event, 0); } free(new_client); return pcmk_ok; } static gboolean cib_async_timeout_handler(gpointer data) { struct timer_rec_s *timer = data; crm_debug("Async call %d timed out after %ds", timer->call_id, timer->timeout); cib_native_callback(timer->cib, NULL, timer->call_id, -ETIME); // We remove the handler in remove_cib_op_callback() return G_SOURCE_CONTINUE; } static gboolean cib_client_register_callback_full(cib_t *cib, int call_id, int timeout, gboolean only_success, void *user_data, const char *callback_name, void (*callback)(xmlNode *, int, int, xmlNode *, void *), void (*free_func)(void *)) { cib_callback_client_t *blob = NULL; if (call_id < 0) { if (only_success == FALSE) { callback(NULL, call_id, call_id, NULL, user_data); } else { crm_warn("CIB call failed: %s", pcmk_strerror(call_id)); } if (user_data && free_func) { free_func(user_data); } return FALSE; } blob = calloc(1, sizeof(cib_callback_client_t)); blob->id = callback_name; blob->only_success = only_success; blob->user_data = user_data; blob->callback = callback; blob->free_func = free_func; if (timeout > 0) { struct timer_rec_s *async_timer = NULL; async_timer = calloc(1, sizeof(struct timer_rec_s)); blob->timer = async_timer; async_timer->cib = cib; async_timer->call_id = call_id; async_timer->timeout = timeout * 1000; async_timer->ref = g_timeout_add(async_timer->timeout, cib_async_timeout_handler, async_timer); } crm_trace("Adding callback %s for call %d", callback_name, call_id); pcmk__intkey_table_insert(cib_op_callback_table, call_id, blob); return TRUE; } static gboolean cib_client_register_callback(cib_t *cib, int call_id, int timeout, gboolean only_success, void *user_data, const char *callback_name, void (*callback) (xmlNode *, int, int, xmlNode *, void *)) { return cib_client_register_callback_full(cib, call_id, timeout, only_success, user_data, callback_name, callback, NULL); } static int cib_client_noop(cib_t * cib, int call_options) { op_common(cib); return cib_internal_op(cib, PCMK__CIB_REQUEST_NOOP, NULL, NULL, NULL, NULL, call_options, NULL); } static int cib_client_ping(cib_t * cib, xmlNode ** output_data, int call_options) { op_common(cib); return cib_internal_op(cib, CRM_OP_PING, NULL, NULL, NULL, output_data, call_options, NULL); } static int cib_client_query(cib_t * cib, const char *section, xmlNode ** output_data, int call_options) { return cib->cmds->query_from(cib, NULL, section, output_data, call_options); } static int cib_client_query_from(cib_t * cib, const char *host, const char *section, xmlNode ** output_data, int call_options) { op_common(cib); return cib_internal_op(cib, PCMK__CIB_REQUEST_QUERY, host, section, NULL, output_data, call_options, NULL); } static int is_primary(cib_t *cib) { op_common(cib); return cib_internal_op(cib, PCMK__CIB_REQUEST_IS_PRIMARY, NULL, NULL, NULL, NULL, cib_scope_local|cib_sync_call, NULL); } static int set_secondary(cib_t *cib, int call_options) { op_common(cib); return cib_internal_op(cib, PCMK__CIB_REQUEST_SECONDARY, NULL, NULL, NULL, NULL, call_options, NULL); } static int set_all_secondary(cib_t * cib, int call_options) { return -EPROTONOSUPPORT; } static int set_primary(cib_t *cib, int call_options) { op_common(cib); crm_trace("Adding cib_scope_local to options"); return cib_internal_op(cib, PCMK__CIB_REQUEST_PRIMARY, NULL, NULL, NULL, NULL, call_options|cib_scope_local, NULL); } static int cib_client_bump_epoch(cib_t * cib, int call_options) { op_common(cib); return cib_internal_op(cib, PCMK__CIB_REQUEST_BUMP, NULL, NULL, NULL, NULL, call_options, NULL); } static int cib_client_upgrade(cib_t * cib, int call_options) { op_common(cib); return cib_internal_op(cib, PCMK__CIB_REQUEST_UPGRADE, NULL, NULL, NULL, NULL, call_options, NULL); } static int cib_client_sync(cib_t * cib, const char *section, int call_options) { return cib->cmds->sync_from(cib, NULL, section, call_options); } static int cib_client_sync_from(cib_t * cib, const char *host, const char *section, int call_options) { op_common(cib); return cib_internal_op(cib, PCMK__CIB_REQUEST_SYNC_TO_ALL, host, section, NULL, NULL, call_options, NULL); } static int cib_client_create(cib_t * cib, const char *section, xmlNode * data, int call_options) { op_common(cib); return cib_internal_op(cib, PCMK__CIB_REQUEST_CREATE, NULL, section, data, NULL, call_options, NULL); } static int cib_client_modify(cib_t * cib, const char *section, xmlNode * data, int call_options) { op_common(cib); return cib_internal_op(cib, PCMK__CIB_REQUEST_MODIFY, NULL, section, data, NULL, call_options, NULL); } -static int -cib_client_update(cib_t * cib, const char *section, xmlNode * data, int call_options) -{ - op_common(cib); - return cib_internal_op(cib, PCMK__CIB_REQUEST_MODIFY, NULL, section, data, - NULL, call_options, NULL); -} - static int cib_client_replace(cib_t * cib, const char *section, xmlNode * data, int call_options) { op_common(cib); return cib_internal_op(cib, PCMK__CIB_REQUEST_REPLACE, NULL, section, data, NULL, call_options, NULL); } static int cib_client_delete(cib_t * cib, const char *section, xmlNode * data, int call_options) { op_common(cib); return cib_internal_op(cib, PCMK__CIB_REQUEST_DELETE, NULL, section, data, NULL, call_options, NULL); } static int cib_client_delete_absolute(cib_t * cib, const char *section, xmlNode * data, int call_options) { op_common(cib); return cib_internal_op(cib, PCMK__CIB_REQUEST_ABS_DELETE, NULL, section, data, NULL, call_options, NULL); } static int cib_client_erase(cib_t * cib, xmlNode ** output_data, int call_options) { op_common(cib); return cib_internal_op(cib, PCMK__CIB_REQUEST_ERASE, NULL, NULL, NULL, output_data, call_options, NULL); } static void cib_destroy_op_callback(gpointer data) { cib_callback_client_t *blob = data; if (blob->timer && blob->timer->ref > 0) { g_source_remove(blob->timer->ref); } free(blob->timer); if (blob->user_data && blob->free_func) { blob->free_func(blob->user_data); } free(blob); } static void destroy_op_callback_table(void) { if (cib_op_callback_table != NULL) { g_hash_table_destroy(cib_op_callback_table); cib_op_callback_table = NULL; } } char * get_shadow_file(const char *suffix) { char *cib_home = NULL; char *fullname = NULL; char *name = crm_strdup_printf("shadow.%s", suffix); const char *dir = getenv("CIB_shadow_dir"); if (dir == NULL) { uid_t uid = geteuid(); struct passwd *pwent = getpwuid(uid); const char *user = NULL; if (pwent) { user = pwent->pw_name; } else { user = getenv("USER"); crm_perror(LOG_ERR, "Assuming %s because cannot get user details for user ID %d", (user? user : "unprivileged user"), uid); } if (pcmk__strcase_any_of(user, "root", CRM_DAEMON_USER, NULL)) { dir = CRM_CONFIG_DIR; } else { const char *home = NULL; if ((home = getenv("HOME")) == NULL) { if (pwent) { home = pwent->pw_dir; } } dir = pcmk__get_tmpdir(); if (home && home[0] == '/') { int rc = 0; cib_home = crm_strdup_printf("%s/.cib", home); rc = mkdir(cib_home, 0700); if (rc < 0 && errno != EEXIST) { crm_perror(LOG_ERR, "Couldn't create user-specific shadow directory: %s", cib_home); errno = 0; } else { dir = cib_home; } } } } fullname = crm_strdup_printf("%s/%s", dir, name); free(cib_home); free(name); return fullname; } cib_t * cib_shadow_new(const char *shadow) { cib_t *new_cib = NULL; char *shadow_file = NULL; CRM_CHECK(shadow != NULL, return NULL); shadow_file = get_shadow_file(shadow); new_cib = cib_file_new(shadow_file); free(shadow_file); return new_cib; } /*! * \brief Create a new CIB connection object, ignoring any active shadow CIB * * Create a new live, file, or remote CIB connection object based on the values * of CIB-related environment variables (CIB_file, CIB_port, CIB_server, * CIB_user, and CIB_passwd). The object will not be connected. * * \return Newly allocated CIB connection object * \note The CIB API does not fully support opening multiple CIB connection * objects simultaneously, so the returned object should be treated as a * singleton. */ cib_t * cib_new_no_shadow(void) { const char *shadow = getenv("CIB_shadow"); cib_t *cib = NULL; unsetenv("CIB_shadow"); cib = cib_new(); if (shadow != NULL) { setenv("CIB_shadow", shadow, 1); } return cib; } /*! * \brief Create a new CIB connection object * * Create a new live, remote, file, or shadow file CIB connection object based * on the values of CIB-related environment variables (CIB_shadow, CIB_file, * CIB_port, CIB_server, CIB_user, and CIB_passwd). The object will not be * connected. * * \return Newly allocated CIB connection object * \note The CIB API does not fully support opening multiple CIB connection * objects simultaneously, so the returned object should be treated as a * singleton. */ /* @TODO Ensure all APIs support multiple simultaneous CIB connection objects * (at least cib_free_callbacks() currently does not). */ cib_t * cib_new(void) { const char *value = getenv("CIB_shadow"); int port; if (value && value[0] != 0) { return cib_shadow_new(value); } value = getenv("CIB_file"); if (value) { return cib_file_new(value); } value = getenv("CIB_port"); if (value) { gboolean encrypted = TRUE; const char *server = getenv("CIB_server"); const char *user = getenv("CIB_user"); const char *pass = getenv("CIB_passwd"); /* We don't ensure port is valid (>= 0) because cib_new() currently * can't return NULL in practice, and introducing a NULL return here * could cause core dumps that would previously just cause signon() * failures. */ pcmk__scan_port(value, &port); value = getenv("CIB_encrypted"); if (value && crm_is_true(value) == FALSE) { crm_info("Disabling TLS"); encrypted = FALSE; } if (user == NULL) { user = CRM_DAEMON_USER; crm_info("Defaulting to user: %s", user); } if (server == NULL) { server = "localhost"; crm_info("Defaulting to localhost"); } return cib_remote_new(server, user, pass, port, encrypted); } return cib_native_new(); } /*! * \internal * \brief Create a generic CIB connection instance * * \return Newly allocated and initialized cib_t instance * * \note This is called by each variant's cib_*_new() function before setting * variant-specific values. */ cib_t * cib_new_variant(void) { cib_t *new_cib = NULL; new_cib = calloc(1, sizeof(cib_t)); if (new_cib == NULL) { return NULL; } remove_cib_op_callback(0, TRUE); /* remove all */ new_cib->call_id = 1; new_cib->variant = cib_undefined; new_cib->type = cib_no_connection; new_cib->state = cib_disconnected; new_cib->op_callback = NULL; new_cib->variant_opaque = NULL; new_cib->notify_list = NULL; /* the rest will get filled in by the variant constructor */ new_cib->cmds = calloc(1, sizeof(cib_api_operations_t)); if (new_cib->cmds == NULL) { free(new_cib); return NULL; } new_cib->cmds->set_op_callback = cib_client_set_op_callback; new_cib->cmds->add_notify_callback = cib_client_add_notify_callback; new_cib->cmds->del_notify_callback = cib_client_del_notify_callback; new_cib->cmds->register_callback = cib_client_register_callback; new_cib->cmds->register_callback_full = cib_client_register_callback_full; new_cib->cmds->noop = cib_client_noop; new_cib->cmds->ping = cib_client_ping; new_cib->cmds->query = cib_client_query; new_cib->cmds->sync = cib_client_sync; new_cib->cmds->query_from = cib_client_query_from; new_cib->cmds->sync_from = cib_client_sync_from; new_cib->cmds->is_master = is_primary; // Deprecated method new_cib->cmds->set_primary = set_primary; new_cib->cmds->set_master = set_primary; // Deprecated method new_cib->cmds->set_secondary = set_secondary; new_cib->cmds->set_slave = set_secondary; // Deprecated method new_cib->cmds->set_slave_all = set_all_secondary; // Deprecated method new_cib->cmds->upgrade = cib_client_upgrade; new_cib->cmds->bump_epoch = cib_client_bump_epoch; new_cib->cmds->create = cib_client_create; new_cib->cmds->modify = cib_client_modify; - new_cib->cmds->update = cib_client_update; + new_cib->cmds->update = cib_client_modify; new_cib->cmds->replace = cib_client_replace; new_cib->cmds->remove = cib_client_delete; new_cib->cmds->erase = cib_client_erase; new_cib->cmds->delete_absolute = cib_client_delete_absolute; return new_cib; } void cib_free_notify(cib_t *cib) { if (cib) { GList *list = cib->notify_list; while (list != NULL) { cib_notify_client_t *client = g_list_nth_data(list, 0); list = g_list_remove(list, client); free(client); } cib->notify_list = NULL; } } /*! * \brief Free all callbacks for a CIB connection * * \param[in,out] cib CIB connection to clean up */ void cib_free_callbacks(cib_t *cib) { cib_free_notify(cib); destroy_op_callback_table(); } /*! * \brief Free all memory used by CIB connection * * \param[in,out] cib CIB connection to delete */ void cib_delete(cib_t *cib) { cib_free_callbacks(cib); if (cib) { cib->cmds->free(cib); } } void remove_cib_op_callback(int call_id, gboolean all_callbacks) { if (all_callbacks) { destroy_op_callback_table(); cib_op_callback_table = pcmk__intkey_table(cib_destroy_op_callback); } else { pcmk__intkey_table_remove(cib_op_callback_table, call_id); } } int num_cib_op_callbacks(void) { if (cib_op_callback_table == NULL) { return 0; } return g_hash_table_size(cib_op_callback_table); } static void cib_dump_pending_op(gpointer key, gpointer value, gpointer user_data) { int call = GPOINTER_TO_INT(key); cib_callback_client_t *blob = value; crm_debug("Call %d (%s): pending", call, pcmk__s(blob->id, "without ID")); } void cib_dump_pending_callbacks(void) { if (cib_op_callback_table == NULL) { return; } return g_hash_table_foreach(cib_op_callback_table, cib_dump_pending_op, NULL); } cib_callback_client_t* cib__lookup_id (int call_id) { return pcmk__intkey_table_lookup(cib_op_callback_table, call_id); } diff --git a/tools/crm_resource_ban.c b/tools/crm_resource_ban.c index c1448e5b52..b1edac8504 100644 --- a/tools/crm_resource_ban.c +++ b/tools/crm_resource_ban.c @@ -1,503 +1,505 @@ /* * 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 General Public License version 2 * or later (GPLv2+) WITHOUT ANY WARRANTY. */ #include #include static char * parse_cli_lifetime(pcmk__output_t *out, const char *move_lifetime) { char *later_s = NULL; crm_time_t *now = NULL; crm_time_t *later = NULL; crm_time_t *duration = NULL; if (move_lifetime == NULL) { return NULL; } duration = crm_time_parse_duration(move_lifetime); if (duration == NULL) { out->err(out, "Invalid duration specified: %s\n" "Please refer to https://en.wikipedia.org/wiki/ISO_8601#Durations " "for examples of valid durations", move_lifetime); return NULL; } now = crm_time_new(NULL); later = crm_time_add(now, duration); if (later == NULL) { out->err(out, "Unable to add %s to current time\n" "Please report to " PACKAGE_BUGREPORT " as possible bug", move_lifetime); crm_time_free(now); crm_time_free(duration); return NULL; } crm_time_log(LOG_INFO, "now ", now, crm_time_log_date | crm_time_log_timeofday | crm_time_log_with_timezone); crm_time_log(LOG_INFO, "later ", later, crm_time_log_date | crm_time_log_timeofday | crm_time_log_with_timezone); crm_time_log(LOG_INFO, "duration", duration, crm_time_log_date | crm_time_log_timeofday); later_s = crm_time_as_string(later, crm_time_log_date | crm_time_log_timeofday | crm_time_log_with_timezone); out->info(out, "Migration will take effect until: %s", later_s); crm_time_free(duration); crm_time_free(later); crm_time_free(now); return later_s; } static const char * promoted_role_name(void) { /* This is a judgment call for what string to use. @TODO Ideally we'd * use the legacy string if the DC only supports that, and the new one * otherwise. Basing it on --enable-compat-2.0 is a decent guess. */ #ifdef PCMK__COMPAT_2_0 return RSC_ROLE_PROMOTED_LEGACY_S; #else return RSC_ROLE_PROMOTED_S; #endif } // \return Standard Pacemaker return code int cli_resource_ban(pcmk__output_t *out, const char *rsc_id, const char *host, const char *move_lifetime, GList *allnodes, cib_t * cib_conn, int cib_options, gboolean promoted_role_only) { char *later_s = NULL; int rc = pcmk_rc_ok; xmlNode *fragment = NULL; xmlNode *location = NULL; if(host == NULL) { GList *n = allnodes; for(; n && rc == pcmk_rc_ok; n = n->next) { pe_node_t *target = n->data; rc = cli_resource_ban(out, rsc_id, target->details->uname, move_lifetime, NULL, cib_conn, cib_options, promoted_role_only); } return rc; } later_s = parse_cli_lifetime(out, move_lifetime); if(move_lifetime && later_s == NULL) { return EINVAL; } fragment = create_xml_node(NULL, XML_CIB_TAG_CONSTRAINTS); location = create_xml_node(fragment, XML_CONS_TAG_RSC_LOCATION); crm_xml_set_id(location, "cli-ban-%s-on-%s", rsc_id, host); out->info(out, "WARNING: Creating rsc_location constraint '%s' with a " "score of -INFINITY for resource %s on %s.\n\tThis will " "prevent %s from %s on %s until the constraint is removed " "using the clear option or by editing the CIB with an " "appropriate tool\n\tThis will be the case even if %s " "is the last node in the cluster", ID(location), rsc_id, host, rsc_id, (promoted_role_only? "being promoted" : "running"), host, host); crm_xml_add(location, XML_LOC_ATTR_SOURCE, rsc_id); if(promoted_role_only) { crm_xml_add(location, XML_RULE_ATTR_ROLE, promoted_role_name()); } else { crm_xml_add(location, XML_RULE_ATTR_ROLE, RSC_ROLE_STARTED_S); } if (later_s == NULL) { /* Short form */ crm_xml_add(location, XML_CIB_TAG_NODE, host); crm_xml_add(location, XML_RULE_ATTR_SCORE, CRM_MINUS_INFINITY_S); } else { xmlNode *rule = create_xml_node(location, XML_TAG_RULE); xmlNode *expr = create_xml_node(rule, XML_TAG_EXPRESSION); crm_xml_set_id(rule, "cli-ban-%s-on-%s-rule", rsc_id, host); crm_xml_add(rule, XML_RULE_ATTR_SCORE, CRM_MINUS_INFINITY_S); crm_xml_add(rule, XML_RULE_ATTR_BOOLEAN_OP, "and"); crm_xml_set_id(expr, "cli-ban-%s-on-%s-expr", rsc_id, host); crm_xml_add(expr, XML_EXPR_ATTR_ATTRIBUTE, CRM_ATTR_UNAME); crm_xml_add(expr, XML_EXPR_ATTR_OPERATION, "eq"); crm_xml_add(expr, XML_EXPR_ATTR_VALUE, host); crm_xml_add(expr, XML_EXPR_ATTR_TYPE, "string"); expr = create_xml_node(rule, "date_expression"); crm_xml_set_id(expr, "cli-ban-%s-on-%s-lifetime", rsc_id, host); crm_xml_add(expr, "operation", "lt"); crm_xml_add(expr, "end", later_s); } crm_log_xml_notice(fragment, "Modify"); - rc = cib_conn->cmds->update(cib_conn, XML_CIB_TAG_CONSTRAINTS, fragment, cib_options); + rc = cib_conn->cmds->modify(cib_conn, XML_CIB_TAG_CONSTRAINTS, fragment, + cib_options); rc = pcmk_legacy2rc(rc); free_xml(fragment); free(later_s); return rc; } // \return Standard Pacemaker return code int cli_resource_prefer(pcmk__output_t *out,const char *rsc_id, const char *host, const char *move_lifetime, cib_t * cib_conn, int cib_options, gboolean promoted_role_only) { char *later_s = parse_cli_lifetime(out, move_lifetime); int rc = pcmk_rc_ok; xmlNode *location = NULL; xmlNode *fragment = NULL; if(move_lifetime && later_s == NULL) { return EINVAL; } if(cib_conn == NULL) { free(later_s); return ENOTCONN; } fragment = create_xml_node(NULL, XML_CIB_TAG_CONSTRAINTS); location = create_xml_node(fragment, XML_CONS_TAG_RSC_LOCATION); crm_xml_set_id(location, "cli-prefer-%s", rsc_id); crm_xml_add(location, XML_LOC_ATTR_SOURCE, rsc_id); if(promoted_role_only) { crm_xml_add(location, XML_RULE_ATTR_ROLE, promoted_role_name()); } else { crm_xml_add(location, XML_RULE_ATTR_ROLE, RSC_ROLE_STARTED_S); } if (later_s == NULL) { /* Short form */ crm_xml_add(location, XML_CIB_TAG_NODE, host); crm_xml_add(location, XML_RULE_ATTR_SCORE, CRM_INFINITY_S); } else { xmlNode *rule = create_xml_node(location, XML_TAG_RULE); xmlNode *expr = create_xml_node(rule, XML_TAG_EXPRESSION); crm_xml_set_id(rule, "cli-prefer-rule-%s", rsc_id); crm_xml_add(rule, XML_RULE_ATTR_SCORE, CRM_INFINITY_S); crm_xml_add(rule, XML_RULE_ATTR_BOOLEAN_OP, "and"); crm_xml_set_id(expr, "cli-prefer-expr-%s", rsc_id); crm_xml_add(expr, XML_EXPR_ATTR_ATTRIBUTE, CRM_ATTR_UNAME); crm_xml_add(expr, XML_EXPR_ATTR_OPERATION, "eq"); crm_xml_add(expr, XML_EXPR_ATTR_VALUE, host); crm_xml_add(expr, XML_EXPR_ATTR_TYPE, "string"); expr = create_xml_node(rule, "date_expression"); crm_xml_set_id(expr, "cli-prefer-lifetime-end-%s", rsc_id); crm_xml_add(expr, "operation", "lt"); crm_xml_add(expr, "end", later_s); } crm_log_xml_info(fragment, "Modify"); - rc = cib_conn->cmds->update(cib_conn, XML_CIB_TAG_CONSTRAINTS, fragment, cib_options); + rc = cib_conn->cmds->modify(cib_conn, XML_CIB_TAG_CONSTRAINTS, fragment, + cib_options); rc = pcmk_legacy2rc(rc); free_xml(fragment); free(later_s); return rc; } /* Nodes can be specified two different ways in the CIB, so we have two different * functions to try clearing out any constraints on them: * * (1) The node could be given by attribute=/value= in an expression XML node. * That's what resource_clear_node_in_expr handles. That XML looks like this: * * * * * * * * * (2) The mode could be given by node= in an rsc_location XML node. That's * what resource_clear_node_in_location handles. That XML looks like this: * * * * \return Standard Pacemaker return code */ static int resource_clear_node_in_expr(const char *rsc_id, const char *host, cib_t * cib_conn, int cib_options) { int rc = pcmk_rc_ok; char *xpath_string = NULL; #define XPATH_FMT \ "//" XML_CONS_TAG_RSC_LOCATION "[@" XML_ATTR_ID "='cli-prefer-%s']" \ "[" XML_TAG_RULE \ "[@" XML_ATTR_ID "='cli-prefer-rule-%s']" \ "/" XML_TAG_EXPRESSION \ "[@" XML_EXPR_ATTR_ATTRIBUTE "='#uname' " \ "and @" XML_EXPR_ATTR_VALUE "='%s']" \ "]" xpath_string = crm_strdup_printf(XPATH_FMT, rsc_id, rsc_id, host); rc = cib_conn->cmds->remove(cib_conn, xpath_string, NULL, cib_xpath | cib_options); if (rc == -ENXIO) { rc = pcmk_rc_ok; } else { rc = pcmk_legacy2rc(rc); } free(xpath_string); return rc; } // \return Standard Pacemaker return code static int resource_clear_node_in_location(const char *rsc_id, const char *host, cib_t * cib_conn, int cib_options, bool clear_ban_constraints, gboolean force) { int rc = pcmk_rc_ok; xmlNode *fragment = NULL; xmlNode *location = NULL; fragment = create_xml_node(NULL, XML_CIB_TAG_CONSTRAINTS); if (clear_ban_constraints == TRUE) { location = create_xml_node(fragment, XML_CONS_TAG_RSC_LOCATION); crm_xml_set_id(location, "cli-ban-%s-on-%s", rsc_id, host); } location = create_xml_node(fragment, XML_CONS_TAG_RSC_LOCATION); crm_xml_set_id(location, "cli-prefer-%s", rsc_id); if (force == FALSE) { crm_xml_add(location, XML_CIB_TAG_NODE, host); } crm_log_xml_info(fragment, "Delete"); rc = cib_conn->cmds->remove(cib_conn, XML_CIB_TAG_CONSTRAINTS, fragment, cib_options); if (rc == -ENXIO) { rc = pcmk_rc_ok; } else { rc = pcmk_legacy2rc(rc); } free_xml(fragment); return rc; } // \return Standard Pacemaker return code int cli_resource_clear(const char *rsc_id, const char *host, GList *allnodes, cib_t * cib_conn, int cib_options, bool clear_ban_constraints, gboolean force) { int rc = pcmk_rc_ok; if(cib_conn == NULL) { return ENOTCONN; } if (host) { rc = resource_clear_node_in_expr(rsc_id, host, cib_conn, cib_options); /* rc does not tell us whether the previous operation did anything, only * whether it failed or not. Thus, as long as it did not fail, we need * to try the second clear method. */ if (rc == pcmk_rc_ok) { rc = resource_clear_node_in_location(rsc_id, host, cib_conn, cib_options, clear_ban_constraints, force); } } else { GList *n = allnodes; /* Iterate over all nodes, attempting to clear the constraint from each. * On the first error, abort. */ for(; n; n = n->next) { pe_node_t *target = n->data; rc = cli_resource_clear(rsc_id, target->details->uname, NULL, cib_conn, cib_options, clear_ban_constraints, force); if (rc != pcmk_rc_ok) { break; } } } return rc; } static void build_clear_xpath_string(GString *buf, const xmlNode *constraint_node, const char *rsc, const char *node, bool promoted_role_only) { const char *cons_id = ID(constraint_node); const char *cons_rsc = crm_element_value(constraint_node, XML_LOC_ATTR_SOURCE); GString *rsc_role_substr = NULL; CRM_ASSERT(buf != NULL); g_string_truncate(buf, 0); if (!pcmk__starts_with(cons_id, "cli-ban-") && !pcmk__starts_with(cons_id, "cli-prefer-")) { return; } g_string_append(buf, "//" XML_CONS_TAG_RSC_LOCATION); if ((node != NULL) || (rsc != NULL) || promoted_role_only) { g_string_append_c(buf, '['); if (node != NULL) { pcmk__g_strcat(buf, "@" XML_CIB_TAG_NODE "='", node, "'", NULL); if (promoted_role_only || (rsc != NULL)) { g_string_append(buf, " and "); } } if ((rsc != NULL) && promoted_role_only) { rsc_role_substr = g_string_sized_new(64); pcmk__g_strcat(rsc_role_substr, "@" XML_LOC_ATTR_SOURCE "='", rsc, "' " "and @" XML_RULE_ATTR_ROLE "='", promoted_role_name(), "'", NULL); } else if (rsc != NULL) { rsc_role_substr = g_string_sized_new(64); pcmk__g_strcat(rsc_role_substr, "@" XML_LOC_ATTR_SOURCE "='", rsc, "'", NULL); } else if (promoted_role_only) { rsc_role_substr = g_string_sized_new(64); pcmk__g_strcat(rsc_role_substr, "@" XML_RULE_ATTR_ROLE "='", promoted_role_name(), "'", NULL); } if (rsc_role_substr != NULL) { g_string_append(buf, rsc_role_substr->str); } g_string_append_c(buf, ']'); } if (node != NULL) { g_string_append(buf, "|//" XML_CONS_TAG_RSC_LOCATION); if (rsc_role_substr != NULL) { pcmk__g_strcat(buf, "[", rsc_role_substr, "]", NULL); } pcmk__g_strcat(buf, "/" XML_TAG_RULE "[" XML_TAG_EXPRESSION "[@" XML_EXPR_ATTR_ATTRIBUTE "='" CRM_ATTR_UNAME "' " "and @" XML_EXPR_ATTR_VALUE "='", node, "']]", NULL); } g_string_append(buf, "//" PCMK_XE_DATE_EXPRESSION "[@" XML_ATTR_ID "='"); if (pcmk__starts_with(cons_id, "cli-ban-")) { pcmk__g_strcat(buf, cons_id, "-lifetime']", NULL); } else { // starts with "cli-prefer-" pcmk__g_strcat(buf, "cli-prefer-lifetime-end-", cons_rsc, "']", NULL); } if (rsc_role_substr != NULL) { g_string_free(rsc_role_substr, TRUE); } } // \return Standard Pacemaker return code int cli_resource_clear_all_expired(xmlNode *root, cib_t *cib_conn, int cib_options, const char *rsc, const char *node, gboolean promoted_role_only) { GString *buf = NULL; xmlXPathObject *xpathObj = NULL; xmlNode *cib_constraints = NULL; crm_time_t *now = crm_time_new(NULL); int i; int rc = pcmk_rc_ok; cib_constraints = pcmk_find_cib_element(root, XML_CIB_TAG_CONSTRAINTS); xpathObj = xpath_search(cib_constraints, "//" XML_CONS_TAG_RSC_LOCATION); for (i = 0; i < numXpathResults(xpathObj); i++) { xmlNode *constraint_node = getXpathResult(xpathObj, i); xmlNode *date_expr_node = NULL; crm_time_t *end = NULL; if (buf == NULL) { buf = g_string_sized_new(1024); } build_clear_xpath_string(buf, constraint_node, rsc, node, promoted_role_only); if (buf->len == 0) { continue; } date_expr_node = get_xpath_object((const char *) buf->str, constraint_node, LOG_DEBUG); if (date_expr_node == NULL) { continue; } /* And then finally, see if the date expression is expired. If so, * clear the constraint. */ end = crm_time_new(crm_element_value(date_expr_node, "end")); if (crm_time_compare(now, end) == 1) { xmlNode *fragment = NULL; xmlNode *location = NULL; fragment = create_xml_node(NULL, XML_CIB_TAG_CONSTRAINTS); location = create_xml_node(fragment, XML_CONS_TAG_RSC_LOCATION); crm_xml_set_id(location, "%s", ID(constraint_node)); crm_log_xml_info(fragment, "Delete"); rc = cib_conn->cmds->remove(cib_conn, XML_CIB_TAG_CONSTRAINTS, fragment, cib_options); rc = pcmk_legacy2rc(rc); if (rc != pcmk_rc_ok) { goto done; } free_xml(fragment); } crm_time_free(end); } done: if (buf != NULL) { g_string_free(buf, TRUE); } freeXpathObject(xpathObj); crm_time_free(now); return rc; }