diff --git a/daemons/controld/controld_control.c b/daemons/controld/controld_control.c index 3331801f32..a8caf80466 100644 --- a/daemons/controld/controld_control.c +++ b/daemons/controld/controld_control.c @@ -1,818 +1,834 @@ /* * Copyright 2004-2020 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 qb_ipcs_service_t *ipcs = NULL; #if SUPPORT_COROSYNC extern gboolean crm_connect_corosync(crm_cluster_t * cluster); #endif void crm_shutdown(int nsig); gboolean crm_read_options(gpointer user_data); gboolean fsa_has_quorum = FALSE; crm_trigger_t *fsa_source = NULL; crm_trigger_t *config_read = NULL; bool no_quorum_suicide_escalation = FALSE; bool controld_shutdown_lock_enabled = false; /* A_HA_CONNECT */ void do_ha_control(long long action, enum crmd_fsa_cause cause, enum crmd_fsa_state cur_state, enum crmd_fsa_input current_input, fsa_data_t * msg_data) { gboolean registered = FALSE; static crm_cluster_t *cluster = NULL; if (cluster == NULL) { cluster = calloc(1, sizeof(crm_cluster_t)); } if (action & A_HA_DISCONNECT) { crm_cluster_disconnect(cluster); crm_info("Disconnected from the cluster"); controld_set_fsa_input_flags(R_HA_DISCONNECTED); } if (action & A_HA_CONNECT) { crm_set_status_callback(&peer_update_callback); crm_set_autoreap(FALSE); if (is_corosync_cluster()) { #if SUPPORT_COROSYNC registered = crm_connect_corosync(cluster); #endif } if (registered == TRUE) { controld_election_init(cluster->uname); fsa_our_uname = cluster->uname; fsa_our_uuid = cluster->uuid; if(cluster->uuid == NULL) { crm_err("Could not obtain local uuid"); registered = FALSE; } } if (registered == FALSE) { controld_set_fsa_input_flags(R_HA_DISCONNECTED); register_fsa_error(C_FSA_INTERNAL, I_ERROR, NULL); return; } populate_cib_nodes(node_update_none, __FUNCTION__); controld_clear_fsa_input_flags(R_HA_DISCONNECTED); crm_info("Connected to the cluster"); } if (action & ~(A_HA_CONNECT | A_HA_DISCONNECT)) { crm_err("Unexpected action %s in %s", fsa_action2string(action), __FUNCTION__); } } /* A_SHUTDOWN */ void do_shutdown(long long action, enum crmd_fsa_cause cause, enum crmd_fsa_state cur_state, enum crmd_fsa_input current_input, fsa_data_t * msg_data) { /* just in case */ controld_set_fsa_input_flags(R_SHUTDOWN); controld_disconnect_fencer(FALSE); } /* A_SHUTDOWN_REQ */ void do_shutdown_req(long long action, enum crmd_fsa_cause cause, enum crmd_fsa_state cur_state, enum crmd_fsa_input current_input, fsa_data_t * msg_data) { xmlNode *msg = NULL; controld_set_fsa_input_flags(R_SHUTDOWN); //controld_set_fsa_input_flags(R_STAYDOWN); crm_info("Sending shutdown request to all peers (DC is %s)", (fsa_our_dc? fsa_our_dc : "not set")); msg = create_request(CRM_OP_SHUTDOWN_REQ, NULL, NULL, CRM_SYSTEM_CRMD, CRM_SYSTEM_CRMD, NULL); if (send_cluster_message(NULL, crm_msg_crmd, msg, TRUE) == FALSE) { register_fsa_error(C_FSA_INTERNAL, I_ERROR, NULL); } free_xml(msg); } extern char *max_generation_from; extern xmlNode *max_generation_xml; extern GHashTable *resource_history; extern GHashTable *voted; void crmd_fast_exit(crm_exit_t exit_code) { if (is_set(fsa_input_register, R_STAYDOWN)) { crm_warn("Inhibiting respawn "CRM_XS" remapping exit code %d to %d", exit_code, CRM_EX_FATAL); exit_code = CRM_EX_FATAL; } else if ((exit_code == CRM_EX_OK) && is_set(fsa_input_register, R_IN_RECOVERY)) { crm_err("Could not recover from internal error"); exit_code = CRM_EX_ERROR; } crm_exit(exit_code); } crm_exit_t crmd_exit(crm_exit_t exit_code) { GListPtr gIter = NULL; GMainLoop *mloop = crmd_mainloop; static bool in_progress = FALSE; if (in_progress && (exit_code == CRM_EX_OK)) { crm_debug("Exit is already in progress"); return exit_code; } else if(in_progress) { crm_notice("Error during shutdown process, exiting now with status %d (%s)", exit_code, crm_exit_str(exit_code)); crm_write_blackbox(SIGTRAP, NULL); crmd_fast_exit(exit_code); } in_progress = TRUE; crm_trace("Preparing to exit with status %d (%s)", exit_code, crm_exit_str(exit_code)); /* Suppress secondary errors resulting from us disconnecting everything */ controld_set_fsa_input_flags(R_HA_DISCONNECTED); /* Close all IPC servers and clients to ensure any and all shared memory files are cleaned up */ if(ipcs) { crm_trace("Closing IPC server"); mainloop_del_ipc_server(ipcs); ipcs = NULL; } controld_close_attrd_ipc(); pe_subsystem_free(); controld_disconnect_fencer(TRUE); if ((exit_code == CRM_EX_OK) && (crmd_mainloop == NULL)) { crm_debug("No mainloop detected"); exit_code = CRM_EX_ERROR; } /* On an error, just get out. * * Otherwise, make the effort to have mainloop exit gracefully so * that it (mostly) cleans up after itself and valgrind has less * to report on - allowing real errors stand out */ if (exit_code != CRM_EX_OK) { crm_notice("Forcing immediate exit with status %d (%s)", exit_code, crm_exit_str(exit_code)); crm_write_blackbox(SIGTRAP, NULL); crmd_fast_exit(exit_code); } /* Clean up as much memory as possible for valgrind */ for (gIter = fsa_message_queue; gIter != NULL; gIter = gIter->next) { fsa_data_t *fsa_data = gIter->data; crm_info("Dropping %s: [ state=%s cause=%s origin=%s ]", fsa_input2string(fsa_data->fsa_input), fsa_state2string(fsa_state), fsa_cause2string(fsa_data->fsa_cause), fsa_data->origin); delete_fsa_input(fsa_data); } controld_clear_fsa_input_flags(R_MEMBERSHIP); g_list_free(fsa_message_queue); fsa_message_queue = NULL; metadata_cache_fini(); controld_election_fini(); /* Tear down the CIB manager connection, but don't free it yet -- it could * be used when we drain the mainloop later. */ cib_free_callbacks(fsa_cib_conn); fsa_cib_conn->cmds->signoff(fsa_cib_conn); verify_stopped(fsa_state, LOG_WARNING); controld_clear_fsa_input_flags(R_LRM_CONNECTED); lrm_state_destroy_all(); /* This basically will not work, since mainloop has a reference to it */ mainloop_destroy_trigger(fsa_source); fsa_source = NULL; mainloop_destroy_trigger(config_read); config_read = NULL; mainloop_destroy_trigger(transition_trigger); transition_trigger = NULL; pcmk__client_cleanup(); crm_peer_destroy(); controld_free_fsa_timers(); te_cleanup_stonith_history_sync(NULL, TRUE); controld_free_sched_timer(); free(fsa_our_dc_version); fsa_our_dc_version = NULL; free(fsa_our_uname); fsa_our_uname = NULL; free(fsa_our_uuid); fsa_our_uuid = NULL; free(fsa_our_dc); fsa_our_dc = NULL; free(fsa_cluster_name); fsa_cluster_name = NULL; free(te_uuid); te_uuid = NULL; free(failed_stop_offset); failed_stop_offset = NULL; free(failed_start_offset); failed_start_offset = NULL; free(max_generation_from); max_generation_from = NULL; free_xml(max_generation_xml); max_generation_xml = NULL; mainloop_destroy_signal(SIGPIPE); mainloop_destroy_signal(SIGUSR1); mainloop_destroy_signal(SIGTERM); mainloop_destroy_signal(SIGTRAP); /* leave SIGCHLD engaged as we might still want to drain some service-actions */ if (mloop) { GMainContext *ctx = g_main_loop_get_context(crmd_mainloop); /* Don't re-enter this block */ crmd_mainloop = NULL; /* no signals on final draining anymore */ mainloop_destroy_signal(SIGCHLD); crm_trace("Draining mainloop %d %d", g_main_loop_is_running(mloop), g_main_context_pending(ctx)); { int lpc = 0; while((g_main_context_pending(ctx) && lpc < 10)) { lpc++; crm_trace("Iteration %d", lpc); g_main_context_dispatch(ctx); } } crm_trace("Closing mainloop %d %d", g_main_loop_is_running(mloop), g_main_context_pending(ctx)); g_main_loop_quit(mloop); /* Won't do anything yet, since we're inside it now */ g_main_loop_unref(mloop); } else { mainloop_destroy_signal(SIGCHLD); } cib_delete(fsa_cib_conn); fsa_cib_conn = NULL; throttle_fini(); /* Graceful */ crm_trace("Done preparing for exit with status %d (%s)", exit_code, crm_exit_str(exit_code)); return exit_code; } /* A_EXIT_0, A_EXIT_1 */ void do_exit(long long action, enum crmd_fsa_cause cause, enum crmd_fsa_state cur_state, enum crmd_fsa_input current_input, fsa_data_t * msg_data) { crm_exit_t exit_code = CRM_EX_OK; int log_level = LOG_INFO; const char *exit_type = "gracefully"; if (action & A_EXIT_1) { log_level = LOG_ERR; exit_type = "forcefully"; exit_code = CRM_EX_ERROR; } verify_stopped(cur_state, LOG_ERR); do_crm_log(log_level, "Performing %s - %s exiting the controller", fsa_action2string(action), exit_type); crm_info("[%s] stopped (%d)", crm_system_name, exit_code); crmd_exit(exit_code); } static void sigpipe_ignore(int nsig) { return; } /* A_STARTUP */ void do_startup(long long action, enum crmd_fsa_cause cause, enum crmd_fsa_state cur_state, enum crmd_fsa_input current_input, fsa_data_t * msg_data) { crm_debug("Registering Signal Handlers"); mainloop_add_signal(SIGTERM, crm_shutdown); mainloop_add_signal(SIGPIPE, sigpipe_ignore); fsa_source = mainloop_add_trigger(G_PRIORITY_HIGH, crm_fsa_trigger, NULL); config_read = mainloop_add_trigger(G_PRIORITY_HIGH, crm_read_options, NULL); transition_trigger = mainloop_add_trigger(G_PRIORITY_LOW, te_graph_trigger, NULL); crm_debug("Creating CIB manager and executor objects"); fsa_cib_conn = cib_new(); lrm_state_init_local(); if (controld_init_fsa_timers() == FALSE) { register_fsa_error(C_FSA_INTERNAL, I_ERROR, NULL); } } // \return libqb error code (0 on success, -errno on error) static int32_t accept_controller_client(qb_ipcs_connection_t *c, uid_t uid, gid_t gid) { crm_trace("Accepting new IPC client connection"); if (pcmk__new_client(c, uid, gid) == NULL) { return -EIO; } return 0; } // \return libqb error code (0 on success, -errno on error) static int32_t dispatch_controller_ipc(qb_ipcs_connection_t * c, void *data, size_t size) { uint32_t id = 0; uint32_t flags = 0; pcmk__client_t *client = pcmk__find_client(c); xmlNode *msg = pcmk__client_data2xml(client, data, &id, &flags); pcmk__ipc_send_ack(client, id, flags, "ack"); if (msg == NULL) { return 0; } #if ENABLE_ACL CRM_ASSERT(client->user != NULL); pcmk__update_acl_user(msg, F_CRM_USER, client->user); #endif crm_xml_add(msg, F_CRM_SYS_FROM, client->id); if (controld_authorize_ipc_message(msg, client, NULL)) { crm_trace("Processing IPC message from %s", pcmk__client_name(client)); route_message(C_IPC_MESSAGE, msg); } trigger_fsa(); free_xml(msg); return 0; } static int32_t crmd_ipc_closed(qb_ipcs_connection_t * c) { pcmk__client_t *client = pcmk__find_client(c); if (client) { crm_trace("Disconnecting %sregistered client %s (%p/%p)", (client->userdata? "" : "un"), pcmk__client_name(client), c, client); free(client->userdata); pcmk__free_client(client); trigger_fsa(); } return 0; } static void crmd_ipc_destroy(qb_ipcs_connection_t * c) { crm_trace("Connection %p", c); crmd_ipc_closed(c); } /* A_STOP */ void do_stop(long long action, enum crmd_fsa_cause cause, enum crmd_fsa_state cur_state, enum crmd_fsa_input current_input, fsa_data_t * msg_data) { crm_trace("Closing IPC server"); mainloop_del_ipc_server(ipcs); ipcs = NULL; register_fsa_input(C_FSA_INTERNAL, I_TERMINATE, NULL); } /* A_STARTED */ void do_started(long long action, enum crmd_fsa_cause cause, enum crmd_fsa_state cur_state, enum crmd_fsa_input current_input, fsa_data_t * msg_data) { static struct qb_ipcs_service_handlers crmd_callbacks = { .connection_accept = accept_controller_client, .connection_created = NULL, .msg_process = dispatch_controller_ipc, .connection_closed = crmd_ipc_closed, .connection_destroyed = crmd_ipc_destroy }; if (cur_state != S_STARTING) { crm_err("Start cancelled... %s", fsa_state2string(cur_state)); return; } else if (is_set(fsa_input_register, R_MEMBERSHIP) == FALSE) { crm_info("Delaying start, no membership data (%.16llx)", R_MEMBERSHIP); crmd_fsa_stall(TRUE); return; } else if (is_set(fsa_input_register, R_LRM_CONNECTED) == FALSE) { crm_info("Delaying start, not connected to executor (%.16llx)", R_LRM_CONNECTED); crmd_fsa_stall(TRUE); return; } else if (is_set(fsa_input_register, R_CIB_CONNECTED) == FALSE) { crm_info("Delaying start, CIB not connected (%.16llx)", R_CIB_CONNECTED); crmd_fsa_stall(TRUE); return; } else if (is_set(fsa_input_register, R_READ_CONFIG) == FALSE) { crm_info("Delaying start, Config not read (%.16llx)", R_READ_CONFIG); crmd_fsa_stall(TRUE); return; } else if (is_set(fsa_input_register, R_PEER_DATA) == FALSE) { crm_info("Delaying start, No peer data (%.16llx)", R_PEER_DATA); crmd_fsa_stall(TRUE); return; } crm_debug("Init server comms"); ipcs = pcmk__serve_controld_ipc(&crmd_callbacks); if (ipcs == NULL) { crm_err("Failed to create IPC server: shutting down and inhibiting respawn"); register_fsa_error(C_FSA_INTERNAL, I_ERROR, NULL); } else { crm_notice("Pacemaker controller successfully started and accepting connections"); } controld_trigger_fencer_connect(); controld_clear_fsa_input_flags(R_STARTING); register_fsa_input(msg_data->fsa_cause, I_PENDING, NULL); } /* A_RECOVER */ void do_recover(long long action, enum crmd_fsa_cause cause, enum crmd_fsa_state cur_state, enum crmd_fsa_input current_input, fsa_data_t * msg_data) { controld_set_fsa_input_flags(R_IN_RECOVERY); crm_warn("Fast-tracking shutdown in response to errors"); register_fsa_input(C_FSA_INTERNAL, I_TERMINATE, NULL); } static pcmk__cluster_option_t crmd_opts[] = { /* name, old name, type, allowed values, * default value, validator, * short description, * long description */ { "dc-version", NULL, "string", NULL, "none", NULL, "Pacemaker version on cluster node elected Designated Controller (DC)", "Includes a hash which identifies the exact changeset the code was " "built from. Used for diagnostic purposes." }, { "cluster-infrastructure", NULL, "string", NULL, "corosync", NULL, "The messaging stack on which Pacemaker is currently running", "Used for informational and diagnostic purposes." }, { "cluster-name", NULL, "string", NULL, NULL, NULL, "An arbitrary name for the cluster", "This optional value is mostly for users' convenience as desired " "in administration, but may also be used in Pacemaker " "configuration rules via the #cluster-name node attribute, and " "by higher-level tools and resource agents." }, { XML_CONFIG_ATTR_DC_DEADTIME, NULL, "time", NULL, "20s", pcmk__valid_interval_spec, "How long to wait for a response from other nodes during start-up", "The optimal value will depend on the speed and load of your network " "and the type of switches used." }, { XML_CONFIG_ATTR_RECHECK, NULL, "time", "Zero disables polling, while positive values are an interval in seconds" "(unless other units are specified, for example \"5min\")", "15min", pcmk__valid_interval_spec, "Polling interval to recheck cluster state and evaluate rules " "with date specifications", "Pacemaker is primarily event-driven, and looks ahead to know when to " "recheck cluster state for failure timeouts and most time-based " "rules. However, it will also recheck the cluster after this " "amount of inactivity, to evaluate rules with date specifications " "and serve as a fail-safe for certain types of scheduler bugs." }, { "load-threshold", NULL, "percentage", NULL, "80%", pcmk__valid_utilization, "Maximum amount of system load that should be used by cluster nodes", "The cluster will slow down its recovery process when the amount of " "system resources used (currently CPU) approaches this limit", }, { "node-action-limit", NULL, "integer", NULL, "0", pcmk__valid_number, "Maximum number of jobs that can be scheduled per node " "(defaults to 2x cores)" }, { XML_CONFIG_ATTR_FENCE_REACTION, NULL, "string", NULL, "stop", NULL, "How a cluster node should react if notified of its own fencing", "A cluster node may receive notification of its own fencing if fencing " "is misconfigured, or if fabric fencing is in use that doesn't cut " "cluster communication. Allowed values are \"stop\" to attempt to " "immediately stop pacemaker and stay stopped, or \"panic\" to attempt " "to immediately reboot the local node, falling back to stop on failure." }, { XML_CONFIG_ATTR_ELECTION_FAIL, NULL, "time", NULL, "2min", pcmk__valid_interval_spec, "*** Advanced Use Only ***", "Declare an election failed if it is not decided within this much " "time. If you need to adjust this value, it probably indicates " "the presence of a bug." }, { XML_CONFIG_ATTR_FORCE_QUIT, NULL, "time", NULL, "20min", pcmk__valid_interval_spec, "*** Advanced Use Only ***", "Exit immediately if shutdown does not complete within this much " "time. If you need to adjust this value, it probably indicates " "the presence of a bug." }, { "join-integration-timeout", "crmd-integration-timeout", "time", NULL, "3min", pcmk__valid_interval_spec, "*** Advanced Use Only ***", "If you need to adjust this value, it probably indicates " "the presence of a bug." }, { "join-finalization-timeout", "crmd-finalization-timeout", "time", NULL, "30min", pcmk__valid_interval_spec, "*** Advanced Use Only ***", "If you need to adjust this value, it probably indicates " "the presence of a bug." }, { "transition-delay", "crmd-transition-delay", "time", NULL, "0s", pcmk__valid_interval_spec, "*** Advanced Use Only *** Enabling this option will slow down " "cluster recovery under all conditions", "Delay cluster recovery for this much time to allow for additional " "events to occur. Useful if your configuration is sensitive to " "the order in which ping updates arrive." }, { "stonith-watchdog-timeout", NULL, "time", NULL, - NULL, pcmk__valid_sbd_timeout, + "0", pcmk__valid_sbd_timeout, "How long to wait before we can assume nodes are safely down " - "when sbd is in use", - NULL + "when watchdog-based self-fencing via SBD is in use", + "If nonzero, along with `have-watchdog=true` automatically set by the " + "cluster, when fencing is required, watchdog-based self-fencing " + "will be performed via SBD without requiring a fencing resource " + "explicitly configured. " + "If `stonith-watchdog-timeout` is set to a positive value, unseen " + "nodes are assumed to self-fence within this much time. +WARNING:+ " + "It must be ensured that this value is larger than the " + "`SBD_WATCHDOG_TIMEOUT` environment variable on all nodes. " + "Pacemaker verifies the settings individually on all nodes and " + "prevents startup or shuts down if configured wrongly on the fly. " + "It's strongly recommended that `SBD_WATCHDOG_TIMEOUT` is set to " + "the same value on all nodes. " + "If `stonith-watchdog-timeout` is set to a negative value, and " + "`SBD_WATCHDOG_TIMEOUT` is set, twice that value will be used. " + "+WARNING:+ In this case, it's essential (currently not verified by " + "pacemaker) that `SBD_WATCHDOG_TIMEOUT` is set to the same value on " + "all nodes." }, { "stonith-max-attempts", NULL, "integer", NULL, "10", pcmk__valid_positive_number, "How many times fencing can fail before it will no longer be " "immediately re-attempted on a target" }, // Already documented in libpe_status (other values must be kept identical) { "no-quorum-policy", NULL, "enum", "stop, freeze, ignore, demote, suicide", "stop", pcmk__valid_quorum, NULL, NULL }, { XML_CONFIG_ATTR_SHUTDOWN_LOCK, NULL, "boolean", NULL, "false", pcmk__valid_boolean, NULL, NULL }, }; void crmd_metadata(void) { pcmk__print_option_metadata("pacemaker-controld", "1.0", "Pacemaker controller options", "Cluster options used by Pacemaker's " "controller (formerly called crmd)", crmd_opts, DIMOF(crmd_opts)); } static void verify_crmd_options(GHashTable * options) { pcmk__validate_cluster_options(options, crmd_opts, DIMOF(crmd_opts)); } static const char * crmd_pref(GHashTable * options, const char *name) { return pcmk__cluster_option(options, crmd_opts, DIMOF(crmd_opts), name); } static void config_query_callback(xmlNode * msg, int call_id, int rc, xmlNode * output, void *user_data) { const char *value = NULL; GHashTable *config_hash = NULL; crm_time_t *now = crm_time_new(NULL); xmlNode *crmconfig = NULL; xmlNode *alerts = NULL; if (rc != pcmk_ok) { fsa_data_t *msg_data = NULL; crm_err("Local CIB query resulted in an error: %s", pcmk_strerror(rc)); register_fsa_error(C_FSA_INTERNAL, I_ERROR, NULL); if (rc == -EACCES || rc == -pcmk_err_schema_validation) { crm_err("The cluster is mis-configured - shutting down and staying down"); controld_set_fsa_input_flags(R_STAYDOWN); } goto bail; } crmconfig = output; if ((crmconfig) && (crm_element_name(crmconfig)) && (strcmp(crm_element_name(crmconfig), XML_CIB_TAG_CRMCONFIG) != 0)) { crmconfig = first_named_child(crmconfig, XML_CIB_TAG_CRMCONFIG); } if (!crmconfig) { fsa_data_t *msg_data = NULL; crm_err("Local CIB query for " XML_CIB_TAG_CRMCONFIG " section failed"); register_fsa_error(C_FSA_INTERNAL, I_ERROR, NULL); goto bail; } crm_debug("Call %d : Parsing CIB options", call_id); config_hash = crm_str_table_new(); pe_unpack_nvpairs(crmconfig, crmconfig, XML_CIB_TAG_PROPSET, NULL, config_hash, CIB_OPTIONS_FIRST, FALSE, now, NULL); verify_crmd_options(config_hash); value = crmd_pref(config_hash, XML_CONFIG_ATTR_DC_DEADTIME); election_trigger->period_ms = crm_parse_interval_spec(value); value = crmd_pref(config_hash, "node-action-limit"); /* Also checks migration-limit */ throttle_update_job_max(value); value = crmd_pref(config_hash, "load-threshold"); if(value) { throttle_set_load_target(strtof(value, NULL) / 100.0); } value = crmd_pref(config_hash, "no-quorum-policy"); if (pcmk__str_eq(value, "suicide", pcmk__str_casei) && pcmk_locate_sbd()) { no_quorum_suicide_escalation = TRUE; } set_fence_reaction(crmd_pref(config_hash, XML_CONFIG_ATTR_FENCE_REACTION)); value = crmd_pref(config_hash,"stonith-max-attempts"); update_stonith_max_attempts(value); value = crmd_pref(config_hash, XML_CONFIG_ATTR_FORCE_QUIT); shutdown_escalation_timer->period_ms = crm_parse_interval_spec(value); crm_debug("Shutdown escalation occurs if DC has not responded to request in %ums", shutdown_escalation_timer->period_ms); value = crmd_pref(config_hash, XML_CONFIG_ATTR_ELECTION_FAIL); controld_set_election_period(value); value = crmd_pref(config_hash, XML_CONFIG_ATTR_RECHECK); recheck_interval_ms = crm_parse_interval_spec(value); crm_debug("Re-run scheduler after %dms of inactivity", recheck_interval_ms); value = crmd_pref(config_hash, "transition-delay"); transition_timer->period_ms = crm_parse_interval_spec(value); value = crmd_pref(config_hash, "join-integration-timeout"); integration_timer->period_ms = crm_parse_interval_spec(value); value = crmd_pref(config_hash, "join-finalization-timeout"); finalization_timer->period_ms = crm_parse_interval_spec(value); value = crmd_pref(config_hash, XML_CONFIG_ATTR_SHUTDOWN_LOCK); controld_shutdown_lock_enabled = crm_is_true(value); free(fsa_cluster_name); fsa_cluster_name = NULL; value = g_hash_table_lookup(config_hash, "cluster-name"); if (value) { fsa_cluster_name = strdup(value); } alerts = first_named_child(output, XML_CIB_TAG_ALERTS); crmd_unpack_alerts(alerts); controld_set_fsa_input_flags(R_READ_CONFIG); crm_trace("Triggering FSA: %s", __FUNCTION__); mainloop_set_trigger(fsa_source); g_hash_table_destroy(config_hash); bail: crm_time_free(now); } gboolean crm_read_options(gpointer user_data) { int call_id = fsa_cib_conn->cmds->query(fsa_cib_conn, "//" XML_CIB_TAG_CRMCONFIG " | //" XML_CIB_TAG_ALERTS, NULL, cib_xpath | cib_scope_local); fsa_register_cib_callback(call_id, FALSE, NULL, config_query_callback); crm_trace("Querying the CIB... call %d", call_id); return TRUE; } /* A_READCONFIG */ void do_read_config(long long action, enum crmd_fsa_cause cause, enum crmd_fsa_state cur_state, enum crmd_fsa_input current_input, fsa_data_t * msg_data) { throttle_init(); mainloop_set_trigger(config_read); } void crm_shutdown(int nsig) { if ((crmd_mainloop == NULL) || !g_main_loop_is_running(crmd_mainloop)) { crmd_exit(CRM_EX_OK); return; } if (is_set(fsa_input_register, R_SHUTDOWN)) { crm_err("Escalating shutdown"); register_fsa_input_before(C_SHUTDOWN, I_ERROR, NULL); return; } controld_set_fsa_input_flags(R_SHUTDOWN); register_fsa_input(C_SHUTDOWN, I_SHUTDOWN, NULL); if (shutdown_escalation_timer->period_ms == 0) { const char *value = crmd_pref(NULL, XML_CONFIG_ATTR_FORCE_QUIT); shutdown_escalation_timer->period_ms = crm_parse_interval_spec(value); } crm_notice("Initiating controller shutdown sequence " CRM_XS " limit=%ums", shutdown_escalation_timer->period_ms); controld_start_timer(shutdown_escalation_timer); } diff --git a/doc/Pacemaker_Explained/en-US/Ch-Options.txt b/doc/Pacemaker_Explained/en-US/Ch-Options.txt index b158f00765..7775cbc589 100644 --- a/doc/Pacemaker_Explained/en-US/Ch-Options.txt +++ b/doc/Pacemaker_Explained/en-US/Ch-Options.txt @@ -1,484 +1,498 @@ :compat-mode: legacy = Cluster-Wide Configuration = == Configuration Layout == The cluster is defined by the Cluster Information Base (CIB), which uses XML notation. The simplest CIB, an empty one, looks like this: .An empty configuration ====== [source,XML] ------- ------- ====== The empty configuration above contains the major sections that make up a CIB: * +cib+: The entire CIB is enclosed with a +cib+ tag. Certain fundamental settings are defined as attributes of this tag. ** +configuration+: This section -- the primary focus of this document -- contains traditional configuration information such as what resources the cluster serves and the relationships among them. *** +crm_config+: cluster-wide configuration options *** +nodes+: the machines that host the cluster *** +resources+: the services run by the cluster *** +constraints+: indications of how resources should be placed ** +status+: This section contains the history of each resource on each node. Based on this data, the cluster can construct the complete current state of the cluster. The authoritative source for this section is the local executor (pacemaker-execd process) on each cluster node, and the cluster will occasionally repopulate the entire section. For this reason, it is never written to disk, and administrators are advised against modifying it in any way. In this document, configuration settings will be described as 'properties' or 'options' based on how they are defined in the CIB: * Properties are XML attributes of an XML element. * Options are name-value pairs expressed as +nvpair+ child elements of an XML element. Normally, you will use command-line tools that abstract the XML, so the distinction will be unimportant; both properties and options are cluster settings you can tweak. == CIB Properties == Certain settings are defined by CIB properties (that is, attributes of the +cib+ tag) rather than with the rest of the cluster configuration in the +configuration+ section. The reason is simply a matter of parsing. These options are used by the configuration database which is, by design, mostly ignorant of the content it holds. So the decision was made to place them in an easy-to-find location. .CIB Properties [width="95%",cols="2m,<5",options="header",align="center"] |========================================================= |Field |Description | admin_epoch | indexterm:[Configuration Version,Cluster] indexterm:[Cluster,Option,Configuration Version] indexterm:[admin_epoch,Cluster Option] indexterm:[Cluster,Option,admin_epoch] When a node joins the cluster, the cluster performs a check to see which node has the best configuration. It asks the node with the highest (+admin_epoch+, +epoch+, +num_updates+) tuple to replace the configuration on all the nodes -- which makes setting them, and setting them correctly, very important. +admin_epoch+ is never modified by the cluster; you can use this to make the configurations on any inactive nodes obsolete. _Never set this value to zero_. In such cases, the cluster cannot tell the difference between your configuration and the "empty" one used when nothing is found on disk. | epoch | indexterm:[epoch,Cluster Option] indexterm:[Cluster,Option,epoch] The cluster increments this every time the configuration is updated (usually by the administrator). | num_updates | indexterm:[num_updates,Cluster Option] indexterm:[Cluster,Option,num_updates] The cluster increments this every time the configuration or status is updated (usually by the cluster) and resets it to 0 when epoch changes. | validate-with | indexterm:[validate-with,Cluster Option] indexterm:[Cluster,Option,validate-with] Determines the type of XML validation that will be done on the configuration. If set to +none+, the cluster will not verify that updates conform to the DTD (nor reject ones that don't). This option can be useful when operating a mixed-version cluster during an upgrade. |cib-last-written | indexterm:[cib-last-written,Cluster Property] indexterm:[Cluster,Property,cib-last-written] Indicates when the configuration was last written to disk. Maintained by the cluster; for informational purposes only. |have-quorum | indexterm:[have-quorum,Cluster Property] indexterm:[Cluster,Property,have-quorum] Indicates if the cluster has quorum. If false, this may mean that the cluster cannot start resources or fence other nodes (see +no-quorum-policy+ below). Maintained by the cluster. |dc-uuid | indexterm:[dc-uuid,Cluster Property] indexterm:[Cluster,Property,dc-uuid] Indicates which cluster node is the current leader. Used by the cluster when placing resources and determining the order of some events. Maintained by the cluster. |========================================================= [[s-cluster-options]] == Cluster Options == Cluster options, as you might expect, control how the cluster behaves when confronted with certain situations. They are grouped into sets within the +crm_config+ section, and, in advanced configurations, there may be more than one set. (This will be described later in the section on <> where we will show how to have the cluster use different sets of options during working hours than during weekends.) For now, we will describe the simple case where each option is present at most once. You can obtain an up-to-date list of cluster options, including their default values, by running the `man pacemaker-schedulerd` and `man pacemaker-controld` commands. .Cluster Options [width="95%",cols="5m,2,<11",options="header",align="center"] |========================================================= |Option |Default |Description | cluster-name | | indexterm:[cluster-name,Cluster Property] indexterm:[Cluster,Property,cluster-name] An (optional) name for the cluster as a whole. This is mostly for users' convenience for use as desired in administration, but this can be used in the Pacemaker configuration in <> (as the +#cluster-name+ <>). It may also be used by higher-level tools when displaying cluster information, and by certain resource agents (for example, the +ocf:heartbeat:GFS2+ agent stores the cluster name in filesystem meta-data). | dc-version | | indexterm:[dc-version,Cluster Property] indexterm:[Cluster,Property,dc-version] Version of Pacemaker on the cluster's DC. Determined automatically by the cluster. Often includes the hash which identifies the exact Git changeset it was built from. Used for diagnostic purposes. | cluster-infrastructure | | indexterm:[cluster-infrastructure,Cluster Property] indexterm:[Cluster,Property,cluster-infrastructure] The messaging stack on which Pacemaker is currently running. Determined automatically by the cluster. Used for informational and diagnostic purposes. | no-quorum-policy | stop a| indexterm:[no-quorum-policy,Cluster Option] indexterm:[Cluster,Option,no-quorum-policy] What to do when the cluster does not have quorum. Allowed values: * +ignore:+ continue all resource management * +freeze:+ continue resource management, but don't recover resources from nodes not in the affected partition * +stop:+ stop all resources in the affected cluster partition * +demote:+ demote promotable resources and stop all other resources in the affected cluster partition * +suicide:+ fence all nodes in the affected cluster partition | batch-limit | 0 | indexterm:[batch-limit,Cluster Option] indexterm:[Cluster,Option,batch-limit] The maximum number of actions that the cluster may execute in parallel across all nodes. The "correct" value will depend on the speed and load of your network and cluster nodes. If zero, the cluster will impose a dynamically calculated limit only when any node has high load. | migration-limit | -1 | indexterm:[migration-limit,Cluster Option] indexterm:[Cluster,Option,migration-limit] The number of <> actions that the cluster is allowed to execute in parallel on a node. A value of -1 means unlimited. | symmetric-cluster | TRUE | indexterm:[symmetric-cluster,Cluster Option] indexterm:[Cluster,Option,symmetric-cluster] Can all resources run on any node by default? | stop-all-resources | FALSE | indexterm:[stop-all-resources,Cluster Option] indexterm:[Cluster,Option,stop-all-resources] Should the cluster stop all resources? | stop-orphan-resources | TRUE | indexterm:[stop-orphan-resources,Cluster Option] indexterm:[Cluster,Option,stop-orphan-resources] Should deleted resources be stopped? This value takes precedence over +is-managed+ (i.e. even unmanaged resources will be stopped if deleted from the configuration when this value is TRUE). | stop-orphan-actions | TRUE | indexterm:[stop-orphan-actions,Cluster Option] indexterm:[Cluster,Option,stop-orphan-actions] Should deleted actions be cancelled? | start-failure-is-fatal | TRUE | indexterm:[start-failure-is-fatal,Cluster Option] indexterm:[Cluster,Option,start-failure-is-fatal] Should a failure to start a resource on a particular node prevent further start attempts on that node? If FALSE, the cluster will decide whether the same node is still eligible based on the resource's current failure count and +migration-threshold+ (see <>). | enable-startup-probes | TRUE | indexterm:[enable-startup-probes,Cluster Option] indexterm:[Cluster,Option,enable-startup-probes] Should the cluster check for active resources during startup? | maintenance-mode | FALSE | indexterm:[maintenance-mode,Cluster Option] indexterm:[Cluster,Option,maintenance-mode] Should the cluster refrain from monitoring, starting and stopping resources? | stonith-enabled | TRUE | indexterm:[stonith-enabled,Cluster Option] indexterm:[Cluster,Option,stonith-enabled] Should failed nodes and nodes with resources that can't be stopped be shot? If you value your data, set up a STONITH device and enable this. If true, or unset, the cluster will refuse to start resources unless one or more STONITH resources have been configured. If false, unresponsive nodes are immediately assumed to be running no resources, and resource takeover to online nodes starts without any further protection (which means _data loss_ if the unresponsive node still accesses shared storage, for example). See also the +requires+ meta-attribute in <>. | stonith-action | reboot | indexterm:[stonith-action,Cluster Option] indexterm:[Cluster,Option,stonith-action] Action to send to STONITH device. Allowed values are +reboot+ and +off+. The value +poweroff+ is also allowed, but is only used for legacy devices. | stonith-timeout | 60s | indexterm:[stonith-timeout,Cluster Option] indexterm:[Cluster,Option,stonith-timeout] How long to wait for STONITH actions (reboot, on, off) to complete | stonith-max-attempts | 10 | indexterm:[stonith-max-attempts,Cluster Option] indexterm:[Cluster,Option,stonith-max-attempts] How many times fencing can fail for a target before the cluster will no longer immediately re-attempt it. | stonith-watchdog-timeout | 0 | indexterm:[stonith-watchdog-timeout,Cluster Option] indexterm:[Cluster,Option,stonith-watchdog-timeout] -If nonzero, rely on hardware watchdog self-fencing. If positive, assume unseen -nodes self-fence within this much time. If negative, and the -SBD_WATCHDOG_TIMEOUT environment variable is set, use twice that value. +If nonzero, along with `have-watchdog=true` automatically set by the +cluster, when fencing is required, watchdog-based self-fencing +will be performed via SBD without requiring a fencing resource +explicitly configured. +If `stonith-watchdog-timeout` is set to a positive value, unseen +nodes are assumed to self-fence within this much time. +WARNING:+ +It must be ensured that this value is larger than the +`SBD_WATCHDOG_TIMEOUT` environment variable on all nodes. +Pacemaker verifies the settings individually on all nodes and +prevents startup or shuts down if configured wrongly on the fly. +It's strongly recommended that `SBD_WATCHDOG_TIMEOUT` is set to +the same value on all nodes. +If `stonith-watchdog-timeout` is set to a negative value, and +`SBD_WATCHDOG_TIMEOUT` is set, twice that value will be used. ++WARNING:+ In this case, it's essential (currently not verified by +pacemaker) that `SBD_WATCHDOG_TIMEOUT` is set to the same value on +all nodes. | concurrent-fencing | FALSE | indexterm:[concurrent-fencing,Cluster Option] indexterm:[Cluster,Option,concurrent-fencing] Is the cluster allowed to initiate multiple fence actions concurrently? | fence-reaction | stop | indexterm:[fence-reaction,Cluster Option] indexterm:[Cluster,Option,fence-reaction] How should a cluster node react if notified of its own fencing? A cluster node may receive notification of its own fencing if fencing is misconfigured, or if fabric fencing is in use that doesn't cut cluster communication. Allowed values are +stop+ to attempt to immediately stop pacemaker and stay stopped, or +panic+ to attempt to immediately reboot the local node, falling back to stop on failure. The default is likely to be changed to +panic+ in a future release. '(since 2.0.3)' | priority-fencing-delay | 0 | indexterm:[priority-fencing-delay,Cluster Option] indexterm:[Cluster,Option,priority-fencing-delay] Apply specified delay for the fencings that are targeting the lost nodes with the highest total resource priority in case we don't have the majority of the nodes in our cluster partition, so that the more significant nodes potentially win any fencing match, which is especially meaningful under split-brain of 2-node cluster. A promoted resource instance takes the base priority + 1 on calculation if the base priority is not 0. Any static/random delays that are introduced by `pcmk_delay_base/max` configured for the corresponding fencing resources will be added to this delay. This delay should be significantly greater than, safely twice, the maximum `pcmk_delay_base/max`. By default, priority fencing delay is disabled. '(since 2.0.4)' | cluster-delay | 60s | indexterm:[cluster-delay,Cluster Option] indexterm:[Cluster,Option,cluster-delay] Estimated maximum round-trip delay over the network (excluding action execution). If the DC requires an action to be executed on another node, it will consider the action failed if it does not get a response from the other node in this time (after considering the action's own timeout). The "correct" value will depend on the speed and load of your network and cluster nodes. | dc-deadtime | 20s | indexterm:[dc-deadtime,Cluster Option] indexterm:[Cluster,Option,dc-deadtime] How long to wait for a response from other nodes during startup. The "correct" value will depend on the speed/load of your network and the type of switches used. | cluster-ipc-limit | 500 | indexterm:[cluster-ipc-limit,Cluster Option] indexterm:[Cluster,Option,cluster-ipc-limit] The maximum IPC message backlog before one cluster daemon will disconnect another. This is of use in large clusters, for which a good value is the number of resources in the cluster multiplied by the number of nodes. The default of 500 is also the minimum. Raise this if you see "Evicting client" messages for cluster daemon PIDs in the logs. | pe-error-series-max | -1 | indexterm:[pe-error-series-max,Cluster Option] indexterm:[Cluster,Option,pe-error-series-max] The number of PE inputs resulting in ERRORs to save. Used when reporting problems. A value of -1 means unlimited (report all). | pe-warn-series-max | -1 | indexterm:[pe-warn-series-max,Cluster Option] indexterm:[Cluster,Option,pe-warn-series-max] The number of PE inputs resulting in WARNINGs to save. Used when reporting problems. A value of -1 means unlimited (report all). | pe-input-series-max | -1 | indexterm:[pe-input-series-max,Cluster Option] indexterm:[Cluster,Option,pe-input-series-max] The number of "normal" PE inputs to save. Used when reporting problems. A value of -1 means unlimited (report all). | placement-strategy | default | indexterm:[placement-strategy,Cluster Option] indexterm:[Cluster,Option,placement-strategy] How the cluster should allocate resources to nodes (see <>). Allowed values are +default+, +utilization+, +balanced+, and +minimal+. | node-health-strategy | none | indexterm:[node-health-strategy,Cluster Option] indexterm:[Cluster,Option,node-health-strategy] How the cluster should react to node health attributes (see <>). Allowed values are +none+, +migrate-on-red+, +only-green+, +progressive+, and +custom+. | enable-acl | FALSE | indexterm:[enable-acl,Cluster Option] indexterm:[Cluster,Option,enable-acl] Whether access control lists (ACLs) (see <>) can be used to authorize modifications to the CIB. | node-health-base | 0 | indexterm:[node-health-base,Cluster Option] indexterm:[Cluster,Option,node-health-base] The base health score assigned to a node. Only used when +node-health-strategy+ is +progressive+. | node-health-green | 0 | indexterm:[node-health-green,Cluster Option] indexterm:[Cluster,Option,node-health-green] The score to use for a node health attribute whose value is +green+. Only used when +node-health-strategy+ is +progressive+ or +custom+. | node-health-yellow | 0 | indexterm:[node-health-yellow,Cluster Option] indexterm:[Cluster,Option,node-health-yellow] The score to use for a node health attribute whose value is +yellow+. Only used when +node-health-strategy+ is +progressive+ or +custom+. | node-health-red | 0 | indexterm:[node-health-red,Cluster Option] indexterm:[Cluster,Option,node-health-red] The score to use for a node health attribute whose value is +red+. Only used when +node-health-strategy+ is +progressive+ or +custom+. | cluster-recheck-interval | 15min | indexterm:[cluster-recheck-interval,Cluster Option] indexterm:[Cluster,Option,cluster-recheck-interval] Pacemaker is primarily event-driven, and looks ahead to know when to recheck the cluster for failure timeouts and most time-based rules. However, it will also recheck the cluster after this amount of inactivity. This has two goals: rules with +date_spec+ are only guaranteed to be checked this often, and it also serves as a fail-safe for certain classes of scheduler bugs. A value of 0 disables this polling; positive values are a time interval. | shutdown-lock | false | The default of false allows active resources to be recovered elsewhere when their node is cleanly shut down, which is what the vast majority of users will want. However, some users prefer to make resources highly available only for failures, with no recovery for clean shutdowns. If this option is true, resources active on a node when it is cleanly shut down are kept "locked" to that node (not allowed to run elsewhere) until they start again on that node after it rejoins (or for at most shutdown-lock-limit, if set). Stonith resources and Pacemaker Remote connections are never locked. Clone and bundle instances and the master role of promotable clones are currently never locked, though support could be added in a future release. Locks may be manually cleared using the `--refresh` option of `crm_resource` (both the resource and node must be specified; this works with remote nodes if their connection resource's target-role is set to Stopped, but not if Pacemaker Remote is stopped on the remote node without disabling the connection resource). '(since 2.0.4)' indexterm:[shutdown-lock,Cluster Option] indexterm:[Cluster,Option,shutdown-lock] | shutdown-lock-limit | 0 | If shutdown-lock is true, and this is set to a nonzero time duration, locked resources will be allowed to start after this much time has passed since the node shutdown was initiated, even if the node has not rejoined. (This works with remote nodes only if their connection resource's target-role is set to Stopped.) '(since 2.0.4)' indexterm:[shutdown-lock-limit,Cluster Option] indexterm:[Cluster,Option,shutdown-lock-limit] | remove-after-stop | FALSE | indexterm:[remove-after-stop,Cluster Option] indexterm:[Cluster,Option,remove-after-stop] _Advanced Use Only:_ Should the cluster remove resources from the LRM after they are stopped? Values other than the default are, at best, poorly tested and potentially dangerous. | startup-fencing | TRUE | indexterm:[startup-fencing,Cluster Option] indexterm:[Cluster,Option,startup-fencing] _Advanced Use Only:_ Should the cluster shoot unseen nodes? Not using the default is very unsafe! | election-timeout | 2min | indexterm:[election-timeout,Cluster Option] indexterm:[Cluster,Option,election-timeout] _Advanced Use Only:_ If you need to adjust this value, it probably indicates the presence of a bug. | shutdown-escalation | 20min | indexterm:[shutdown-escalation,Cluster Option] indexterm:[Cluster,Option,shutdown-escalation] _Advanced Use Only:_ If you need to adjust this value, it probably indicates the presence of a bug. | join-integration-timeout | 3min | indexterm:[join-integration-timeout,Cluster Option] indexterm:[Cluster,Option,join-integration-timeout] _Advanced Use Only:_ If you need to adjust this value, it probably indicates the presence of a bug. | join-finalization-timeout | 30min | indexterm:[join-finalization-timeout,Cluster Option] indexterm:[Cluster,Option,join-finalization-timeout] _Advanced Use Only:_ If you need to adjust this value, it probably indicates the presence of a bug. | transition-delay | 0s | indexterm:[transition-delay,Cluster Option] indexterm:[Cluster,Option,transition-delay] _Advanced Use Only:_ Delay cluster recovery for the configured interval to allow for additional/related events to occur. Useful if your configuration is sensitive to the order in which ping updates arrive. Enabling this option will slow down cluster recovery under all conditions. |========================================================= diff --git a/doc/sphinx/Pacemaker_Explained/options.rst b/doc/sphinx/Pacemaker_Explained/options.rst index 387ace8845..04eb4f437d 100644 --- a/doc/sphinx/Pacemaker_Explained/options.rst +++ b/doc/sphinx/Pacemaker_Explained/options.rst @@ -1,486 +1,500 @@ Cluster-Wide Configuration -------------------------- .. Convert_to_RST: == Configuration Layout == The cluster is defined by the Cluster Information Base (CIB), which uses XML notation. The simplest CIB, an empty one, looks like this: .An empty configuration ====== [source,XML] ------- ------- ====== The empty configuration above contains the major sections that make up a CIB: * +cib+: The entire CIB is enclosed with a +cib+ tag. Certain fundamental settings are defined as attributes of this tag. ** +configuration+: This section -- the primary focus of this document -- contains traditional configuration information such as what resources the cluster serves and the relationships among them. *** +crm_config+: cluster-wide configuration options *** +nodes+: the machines that host the cluster *** +resources+: the services run by the cluster *** +constraints+: indications of how resources should be placed ** +status+: This section contains the history of each resource on each node. Based on this data, the cluster can construct the complete current state of the cluster. The authoritative source for this section is the local executor (pacemaker-execd process) on each cluster node, and the cluster will occasionally repopulate the entire section. For this reason, it is never written to disk, and administrators are advised against modifying it in any way. In this document, configuration settings will be described as 'properties' or 'options' based on how they are defined in the CIB: * Properties are XML attributes of an XML element. * Options are name-value pairs expressed as +nvpair+ child elements of an XML element. Normally, you will use command-line tools that abstract the XML, so the distinction will be unimportant; both properties and options are cluster settings you can tweak. == CIB Properties == Certain settings are defined by CIB properties (that is, attributes of the +cib+ tag) rather than with the rest of the cluster configuration in the +configuration+ section. The reason is simply a matter of parsing. These options are used by the configuration database which is, by design, mostly ignorant of the content it holds. So the decision was made to place them in an easy-to-find location. .CIB Properties [width="95%",cols="2m,<5",options="header",align="center"] |========================================================= |Field |Description | admin_epoch | indexterm:[Configuration Version,Cluster] indexterm:[Cluster,Option,Configuration Version] indexterm:[admin_epoch,Cluster Option] indexterm:[Cluster,Option,admin_epoch] When a node joins the cluster, the cluster performs a check to see which node has the best configuration. It asks the node with the highest (+admin_epoch+, +epoch+, +num_updates+) tuple to replace the configuration on all the nodes -- which makes setting them, and setting them correctly, very important. +admin_epoch+ is never modified by the cluster; you can use this to make the configurations on any inactive nodes obsolete. _Never set this value to zero_. In such cases, the cluster cannot tell the difference between your configuration and the "empty" one used when nothing is found on disk. | epoch | indexterm:[epoch,Cluster Option] indexterm:[Cluster,Option,epoch] The cluster increments this every time the configuration is updated (usually by the administrator). | num_updates | indexterm:[num_updates,Cluster Option] indexterm:[Cluster,Option,num_updates] The cluster increments this every time the configuration or status is updated (usually by the cluster) and resets it to 0 when epoch changes. | validate-with | indexterm:[validate-with,Cluster Option] indexterm:[Cluster,Option,validate-with] Determines the type of XML validation that will be done on the configuration. If set to +none+, the cluster will not verify that updates conform to the DTD (nor reject ones that don't). This option can be useful when operating a mixed-version cluster during an upgrade. |cib-last-written | indexterm:[cib-last-written,Cluster Property] indexterm:[Cluster,Property,cib-last-written] Indicates when the configuration was last written to disk. Maintained by the cluster; for informational purposes only. |have-quorum | indexterm:[have-quorum,Cluster Property] indexterm:[Cluster,Property,have-quorum] Indicates if the cluster has quorum. If false, this may mean that the cluster cannot start resources or fence other nodes (see +no-quorum-policy+ below). Maintained by the cluster. |dc-uuid | indexterm:[dc-uuid,Cluster Property] indexterm:[Cluster,Property,dc-uuid] Indicates which cluster node is the current leader. Used by the cluster when placing resources and determining the order of some events. Maintained by the cluster. |========================================================= [[s-cluster-options]] == Cluster Options == Cluster options, as you might expect, control how the cluster behaves when confronted with certain situations. They are grouped into sets within the +crm_config+ section, and, in advanced configurations, there may be more than one set. (This will be described later in the section on <> where we will show how to have the cluster use different sets of options during working hours than during weekends.) For now, we will describe the simple case where each option is present at most once. You can obtain an up-to-date list of cluster options, including their default values, by running the `man pacemaker-schedulerd` and `man pacemaker-controld` commands. .Cluster Options [width="95%",cols="5m,2,<11",options="header",align="center"] |========================================================= |Option |Default |Description | cluster-name | | indexterm:[cluster-name,Cluster Property] indexterm:[Cluster,Property,cluster-name] An (optional) name for the cluster as a whole. This is mostly for users' convenience for use as desired in administration, but this can be used in the Pacemaker configuration in <> (as the +#cluster-name+ <>). It may also be used by higher-level tools when displaying cluster information, and by certain resource agents (for example, the +ocf:heartbeat:GFS2+ agent stores the cluster name in filesystem meta-data). | dc-version | | indexterm:[dc-version,Cluster Property] indexterm:[Cluster,Property,dc-version] Version of Pacemaker on the cluster's DC. Determined automatically by the cluster. Often includes the hash which identifies the exact Git changeset it was built from. Used for diagnostic purposes. | cluster-infrastructure | | indexterm:[cluster-infrastructure,Cluster Property] indexterm:[Cluster,Property,cluster-infrastructure] The messaging stack on which Pacemaker is currently running. Determined automatically by the cluster. Used for informational and diagnostic purposes. | no-quorum-policy | stop a| indexterm:[no-quorum-policy,Cluster Option] indexterm:[Cluster,Option,no-quorum-policy] What to do when the cluster does not have quorum. Allowed values: * +ignore:+ continue all resource management * +freeze:+ continue resource management, but don't recover resources from nodes not in the affected partition * +stop:+ stop all resources in the affected cluster partition * +demote:+ demote promotable resources and stop all other resources in the affected cluster partition * +suicide:+ fence all nodes in the affected cluster partition | batch-limit | 0 | indexterm:[batch-limit,Cluster Option] indexterm:[Cluster,Option,batch-limit] The maximum number of actions that the cluster may execute in parallel across all nodes. The "correct" value will depend on the speed and load of your network and cluster nodes. If zero, the cluster will impose a dynamically calculated limit only when any node has high load. | migration-limit | -1 | indexterm:[migration-limit,Cluster Option] indexterm:[Cluster,Option,migration-limit] The number of <> actions that the cluster is allowed to execute in parallel on a node. A value of -1 means unlimited. | symmetric-cluster | TRUE | indexterm:[symmetric-cluster,Cluster Option] indexterm:[Cluster,Option,symmetric-cluster] Can all resources run on any node by default? | stop-all-resources | FALSE | indexterm:[stop-all-resources,Cluster Option] indexterm:[Cluster,Option,stop-all-resources] Should the cluster stop all resources? | stop-orphan-resources | TRUE | indexterm:[stop-orphan-resources,Cluster Option] indexterm:[Cluster,Option,stop-orphan-resources] Should deleted resources be stopped? This value takes precedence over +is-managed+ (i.e. even unmanaged resources will be stopped if deleted from the configuration when this value is TRUE). | stop-orphan-actions | TRUE | indexterm:[stop-orphan-actions,Cluster Option] indexterm:[Cluster,Option,stop-orphan-actions] Should deleted actions be cancelled? | start-failure-is-fatal | TRUE | indexterm:[start-failure-is-fatal,Cluster Option] indexterm:[Cluster,Option,start-failure-is-fatal] Should a failure to start a resource on a particular node prevent further start attempts on that node? If FALSE, the cluster will decide whether the same node is still eligible based on the resource's current failure count and +migration-threshold+ (see <>). | enable-startup-probes | TRUE | indexterm:[enable-startup-probes,Cluster Option] indexterm:[Cluster,Option,enable-startup-probes] Should the cluster check for active resources during startup? | maintenance-mode | FALSE | indexterm:[maintenance-mode,Cluster Option] indexterm:[Cluster,Option,maintenance-mode] Should the cluster refrain from monitoring, starting and stopping resources? | stonith-enabled | TRUE | indexterm:[stonith-enabled,Cluster Option] indexterm:[Cluster,Option,stonith-enabled] Should failed nodes and nodes with resources that can't be stopped be shot? If you value your data, set up a STONITH device and enable this. If true, or unset, the cluster will refuse to start resources unless one or more STONITH resources have been configured. If false, unresponsive nodes are immediately assumed to be running no resources, and resource takeover to online nodes starts without any further protection (which means _data loss_ if the unresponsive node still accesses shared storage, for example). See also the +requires+ meta-attribute in <>. | stonith-action | reboot | indexterm:[stonith-action,Cluster Option] indexterm:[Cluster,Option,stonith-action] Action to send to STONITH device. Allowed values are +reboot+ and +off+. The value +poweroff+ is also allowed, but is only used for legacy devices. | stonith-timeout | 60s | indexterm:[stonith-timeout,Cluster Option] indexterm:[Cluster,Option,stonith-timeout] How long to wait for STONITH actions (reboot, on, off) to complete | stonith-max-attempts | 10 | indexterm:[stonith-max-attempts,Cluster Option] indexterm:[Cluster,Option,stonith-max-attempts] How many times fencing can fail for a target before the cluster will no longer immediately re-attempt it. | stonith-watchdog-timeout | 0 | indexterm:[stonith-watchdog-timeout,Cluster Option] indexterm:[Cluster,Option,stonith-watchdog-timeout] - If nonzero, rely on hardware watchdog self-fencing. If positive, assume unseen - nodes self-fence within this much time. If negative, and the - SBD_WATCHDOG_TIMEOUT environment variable is set, use twice that value. - + If nonzero, along with `have-watchdog=true` automatically set by the + cluster, when fencing is required, watchdog-based self-fencing + will be performed via SBD without requiring a fencing resource + explicitly configured. + If `stonith-watchdog-timeout` is set to a positive value, unseen + nodes are assumed to self-fence within this much time. +WARNING:+ + It must be ensured that this value is larger than the + `SBD_WATCHDOG_TIMEOUT` environment variable on all nodes. + Pacemaker verifies the settings individually on all nodes and + prevents startup or shuts down if configured wrongly on the fly. + It's strongly recommended that `SBD_WATCHDOG_TIMEOUT` is set to + the same value on all nodes. + If `stonith-watchdog-timeout` is set to a negative value, and + `SBD_WATCHDOG_TIMEOUT` is set, twice that value will be used. + +WARNING:+ In this case, it's essential (currently not verified by + pacemaker) that `SBD_WATCHDOG_TIMEOUT` is set to the same value on + all nodes. + | concurrent-fencing | FALSE | indexterm:[concurrent-fencing,Cluster Option] indexterm:[Cluster,Option,concurrent-fencing] Is the cluster allowed to initiate multiple fence actions concurrently? | fence-reaction | stop | indexterm:[fence-reaction,Cluster Option] indexterm:[Cluster,Option,fence-reaction] How should a cluster node react if notified of its own fencing? A cluster node may receive notification of its own fencing if fencing is misconfigured, or if fabric fencing is in use that doesn't cut cluster communication. Allowed values are +stop+ to attempt to immediately stop pacemaker and stay stopped, or +panic+ to attempt to immediately reboot the local node, falling back to stop on failure. The default is likely to be changed to +panic+ in a future release. '(since 2.0.3)' | priority-fencing-delay | 0 | indexterm:[priority-fencing-delay,Cluster Option] indexterm:[Cluster,Option,priority-fencing-delay] Apply specified delay for the fencings that are targeting the lost nodes with the highest total resource priority in case we don't have the majority of the nodes in our cluster partition, so that the more significant nodes potentially win any fencing match, which is especially meaningful under split-brain of 2-node cluster. A promoted resource instance takes the base priority + 1 on calculation if the base priority is not 0. Any static/random delays that are introduced by `pcmk_delay_base/max` configured for the corresponding fencing resources will be added to this delay. This delay should be significantly greater than, safely twice, the maximum `pcmk_delay_base/max`. By default, priority fencing delay is disabled. '(since 2.0.4)' | cluster-delay | 60s | indexterm:[cluster-delay,Cluster Option] indexterm:[Cluster,Option,cluster-delay] Estimated maximum round-trip delay over the network (excluding action execution). If the DC requires an action to be executed on another node, it will consider the action failed if it does not get a response from the other node in this time (after considering the action's own timeout). The "correct" value will depend on the speed and load of your network and cluster nodes. | dc-deadtime | 20s | indexterm:[dc-deadtime,Cluster Option] indexterm:[Cluster,Option,dc-deadtime] How long to wait for a response from other nodes during startup. The "correct" value will depend on the speed/load of your network and the type of switches used. | cluster-ipc-limit | 500 | indexterm:[cluster-ipc-limit,Cluster Option] indexterm:[Cluster,Option,cluster-ipc-limit] The maximum IPC message backlog before one cluster daemon will disconnect another. This is of use in large clusters, for which a good value is the number of resources in the cluster multiplied by the number of nodes. The default of 500 is also the minimum. Raise this if you see "Evicting client" messages for cluster daemon PIDs in the logs. | pe-error-series-max | -1 | indexterm:[pe-error-series-max,Cluster Option] indexterm:[Cluster,Option,pe-error-series-max] The number of PE inputs resulting in ERRORs to save. Used when reporting problems. A value of -1 means unlimited (report all). | pe-warn-series-max | -1 | indexterm:[pe-warn-series-max,Cluster Option] indexterm:[Cluster,Option,pe-warn-series-max] The number of PE inputs resulting in WARNINGs to save. Used when reporting problems. A value of -1 means unlimited (report all). | pe-input-series-max | -1 | indexterm:[pe-input-series-max,Cluster Option] indexterm:[Cluster,Option,pe-input-series-max] The number of "normal" PE inputs to save. Used when reporting problems. A value of -1 means unlimited (report all). | placement-strategy | default | indexterm:[placement-strategy,Cluster Option] indexterm:[Cluster,Option,placement-strategy] How the cluster should allocate resources to nodes (see <>). Allowed values are +default+, +utilization+, +balanced+, and +minimal+. | node-health-strategy | none | indexterm:[node-health-strategy,Cluster Option] indexterm:[Cluster,Option,node-health-strategy] How the cluster should react to node health attributes (see <>). Allowed values are +none+, +migrate-on-red+, +only-green+, +progressive+, and +custom+. | enable-acl | FALSE | indexterm:[enable-acl,Cluster Option] indexterm:[Cluster,Option,enable-acl] Whether access control lists (ACLs) (see <>) can be used to authorize modifications to the CIB. | node-health-base | 0 | indexterm:[node-health-base,Cluster Option] indexterm:[Cluster,Option,node-health-base] The base health score assigned to a node. Only used when +node-health-strategy+ is +progressive+. | node-health-green | 0 | indexterm:[node-health-green,Cluster Option] indexterm:[Cluster,Option,node-health-green] The score to use for a node health attribute whose value is +green+. Only used when +node-health-strategy+ is +progressive+ or +custom+. | node-health-yellow | 0 | indexterm:[node-health-yellow,Cluster Option] indexterm:[Cluster,Option,node-health-yellow] The score to use for a node health attribute whose value is +yellow+. Only used when +node-health-strategy+ is +progressive+ or +custom+. | node-health-red | 0 | indexterm:[node-health-red,Cluster Option] indexterm:[Cluster,Option,node-health-red] The score to use for a node health attribute whose value is +red+. Only used when +node-health-strategy+ is +progressive+ or +custom+. | cluster-recheck-interval | 15min | indexterm:[cluster-recheck-interval,Cluster Option] indexterm:[Cluster,Option,cluster-recheck-interval] Pacemaker is primarily event-driven, and looks ahead to know when to recheck the cluster for failure timeouts and most time-based rules. However, it will also recheck the cluster after this amount of inactivity. This has two goals: rules with +date_spec+ are only guaranteed to be checked this often, and it also serves as a fail-safe for certain classes of scheduler bugs. A value of 0 disables this polling; positive values are a time interval. | shutdown-lock | false | The default of false allows active resources to be recovered elsewhere when their node is cleanly shut down, which is what the vast majority of users will want. However, some users prefer to make resources highly available only for failures, with no recovery for clean shutdowns. If this option is true, resources active on a node when it is cleanly shut down are kept "locked" to that node (not allowed to run elsewhere) until they start again on that node after it rejoins (or for at most shutdown-lock-limit, if set). Stonith resources and Pacemaker Remote connections are never locked. Clone and bundle instances and the master role of promotable clones are currently never locked, though support could be added in a future release. Locks may be manually cleared using the `--refresh` option of `crm_resource` (both the resource and node must be specified; this works with remote nodes if their connection resource's target-role is set to Stopped, but not if Pacemaker Remote is stopped on the remote node without disabling the connection resource). '(since 2.0.4)' indexterm:[shutdown-lock,Cluster Option] indexterm:[Cluster,Option,shutdown-lock] | shutdown-lock-limit | 0 | If shutdown-lock is true, and this is set to a nonzero time duration, locked resources will be allowed to start after this much time has passed since the node shutdown was initiated, even if the node has not rejoined. (This works with remote nodes only if their connection resource's target-role is set to Stopped.) '(since 2.0.4)' indexterm:[shutdown-lock-limit,Cluster Option] indexterm:[Cluster,Option,shutdown-lock-limit] | remove-after-stop | FALSE | indexterm:[remove-after-stop,Cluster Option] indexterm:[Cluster,Option,remove-after-stop] _Advanced Use Only:_ Should the cluster remove resources from the LRM after they are stopped? Values other than the default are, at best, poorly tested and potentially dangerous. | startup-fencing | TRUE | indexterm:[startup-fencing,Cluster Option] indexterm:[Cluster,Option,startup-fencing] _Advanced Use Only:_ Should the cluster shoot unseen nodes? Not using the default is very unsafe! | election-timeout | 2min | indexterm:[election-timeout,Cluster Option] indexterm:[Cluster,Option,election-timeout] _Advanced Use Only:_ If you need to adjust this value, it probably indicates the presence of a bug. | shutdown-escalation | 20min | indexterm:[shutdown-escalation,Cluster Option] indexterm:[Cluster,Option,shutdown-escalation] _Advanced Use Only:_ If you need to adjust this value, it probably indicates the presence of a bug. | join-integration-timeout | 3min | indexterm:[join-integration-timeout,Cluster Option] indexterm:[Cluster,Option,join-integration-timeout] _Advanced Use Only:_ If you need to adjust this value, it probably indicates the presence of a bug. | join-finalization-timeout | 30min | indexterm:[join-finalization-timeout,Cluster Option] indexterm:[Cluster,Option,join-finalization-timeout] _Advanced Use Only:_ If you need to adjust this value, it probably indicates the presence of a bug. | transition-delay | 0s | indexterm:[transition-delay,Cluster Option] indexterm:[Cluster,Option,transition-delay] _Advanced Use Only:_ Delay cluster recovery for the configured interval to allow for additional/related events to occur. Useful if your configuration is sensitive to the order in which ping updates arrive. Enabling this option will slow down cluster recovery under all conditions. |=========================================================