diff --git a/crmd/control.c b/crmd/control.c index f4823b97c4..af9c2c2f22 100644 --- a/crmd/control.c +++ b/crmd/control.c @@ -1,1163 +1,1163 @@ /* * Copyright (C) 2004 Andrew Beekhof * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public * License as published by the Free Software Foundation; either * version 2 of the License, or (at your option) any later version. * * This software is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * General Public License for more details. * * You should have received a copy of the GNU General Public * License along with this library; if not, write to the Free Software * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include qb_ipcs_service_t *ipcs = NULL; extern gboolean crm_connect_corosync(crm_cluster_t * cluster); extern void crmd_ha_connection_destroy(gpointer user_data); 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; static gboolean election_timeout_popped(gpointer data) { /* Not everyone voted */ crm_info("Election failed: Declaring ourselves the winner"); register_fsa_input(C_TIMER_POPPED, I_ELECTION_DC, NULL); return 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"); set_bit(fsa_input_register, R_HA_DISCONNECTED); } if (action & A_HA_CONNECT) { crm_set_status_callback(&peer_update_callback); crm_set_autoreap(FALSE); if (is_openais_cluster()) { #if SUPPORT_COROSYNC registered = crm_connect_corosync(cluster); #endif } else if (is_heartbeat_cluster()) { #if SUPPORT_HEARTBEAT cluster->destroy = crmd_ha_connection_destroy; cluster->hb_dispatch = crmd_ha_msg_callback; registered = crm_cluster_connect(cluster); fsa_cluster_conn = cluster->hb_conn; crm_trace("Be informed of Node Status changes"); if (registered && fsa_cluster_conn->llc_ops->set_nstatus_callback(fsa_cluster_conn, crmd_ha_status_callback, fsa_cluster_conn) != HA_OK) { crm_err("Cannot set nstatus callback: %s", fsa_cluster_conn->llc_ops->errmsg(fsa_cluster_conn)); registered = FALSE; } crm_trace("Be informed of CRM Client Status changes"); if (registered && fsa_cluster_conn->llc_ops->set_cstatus_callback(fsa_cluster_conn, crmd_client_status_callback, fsa_cluster_conn) != HA_OK) { crm_err("Cannot set cstatus callback: %s", fsa_cluster_conn->llc_ops->errmsg(fsa_cluster_conn)); registered = FALSE; } if (registered) { crm_trace("Requesting an initial dump of CRMD client_status"); fsa_cluster_conn->llc_ops->client_status(fsa_cluster_conn, NULL, CRM_SYSTEM_CRMD, -1); } #endif } fsa_election = election_init(NULL, cluster->uname, 60000/*60s*/, election_timeout_popped); 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) { set_bit(fsa_input_register, R_HA_DISCONNECTED); register_fsa_error(C_FSA_INTERNAL, I_ERROR, NULL); return; } populate_cib_nodes(node_update_none, __FUNCTION__); clear_bit(fsa_input_register, 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__); } } static bool need_spawn_pengine_from_crmd(void) { static int result = -1; if (result != -1) return result; if (!is_heartbeat_cluster()) { result = 0; return result; } /* NULL, or "strange" value: rather spawn from here. */ result = TRUE; crm_str_to_boolean(daemon_option("crmd_spawns_pengine"), &result); return result; } /* 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 */ set_bit(fsa_input_register, R_SHUTDOWN); if (need_spawn_pengine_from_crmd()) { if (is_set(fsa_input_register, pe_subsystem->flag_connected)) { crm_info("Terminating the %s", pe_subsystem->name); if (stop_subsystem(pe_subsystem, TRUE) == FALSE) { /* its gone... */ crm_err("Faking %s exit", pe_subsystem->name); clear_bit(fsa_input_register, pe_subsystem->flag_connected); } else { crm_info("Waiting for subsystems to exit"); crmd_fsa_stall(FALSE); } } crm_info("All subsystems stopped, continuing"); } if (stonith_api) { /* Prevent it from coming up again */ clear_bit(fsa_input_register, R_ST_REQUIRED); crm_info("Disconnecting STONITH..."); stonith_api->cmds->disconnect(stonith_api); } } /* 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; set_bit(fsa_input_register, R_SHUTDOWN); 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); /* set_bit(fsa_input_register, R_STAYDOWN); */ if (send_cluster_message(NULL, crm_msg_crmd, msg, TRUE) == FALSE) { register_fsa_error(C_FSA_INTERNAL, I_ERROR, NULL); } free_xml(msg); } extern crm_ipc_t *attrd_ipc; extern char *max_generation_from; extern xmlNode *max_generation_xml; extern GHashTable *resource_history; extern GHashTable *voted; extern GHashTable *metadata_hash; extern char *te_client_id; void log_connected_client(gpointer key, gpointer value, gpointer user_data); void log_connected_client(gpointer key, gpointer value, gpointer user_data) { crm_client_t *client = value; crm_err("%s is still connected at exit", crm_client_name(client)); } int crmd_fast_exit(int rc) { if (is_set(fsa_input_register, R_STAYDOWN)) { crm_warn("Inhibiting respawn "CRM_XS" remapping exit code %d to %d", rc, DAEMON_RESPAWN_STOP); rc = DAEMON_RESPAWN_STOP; } if (rc == pcmk_ok && is_set(fsa_input_register, R_IN_RECOVERY)) { crm_err("Could not recover from internal error"); rc = pcmk_err_generic; } return crm_exit(rc); } int crmd_exit(int rc) { GListPtr gIter = NULL; GMainLoop *mloop = crmd_mainloop; static bool in_progress = FALSE; if(in_progress && rc == 0) { crm_debug("Exit is already in progress"); return rc; } else if(in_progress) { crm_notice("Error during shutdown process, terminating now with status %d: %s", rc, pcmk_strerror(rc)); crm_write_blackbox(SIGTRAP, NULL); crmd_fast_exit(rc); } in_progress = TRUE; crm_trace("Preparing to exit: %d", rc); /* Suppress secondary errors resulting from us disconnecting everything */ set_bit(fsa_input_register, 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; } if (attrd_ipc) { crm_trace("Closing attrd connection"); crm_ipc_close(attrd_ipc); crm_ipc_destroy(attrd_ipc); attrd_ipc = NULL; } if (pe_subsystem && pe_subsystem->client && pe_subsystem->client->ipcs) { crm_trace("Disconnecting Policy Engine"); qb_ipcs_disconnect(pe_subsystem->client->ipcs); } if(stonith_api) { crm_trace("Disconnecting fencing API"); clear_bit(fsa_input_register, R_ST_REQUIRED); stonith_api->cmds->free(stonith_api); stonith_api = NULL; } if (rc == pcmk_ok && crmd_mainloop == NULL) { crm_debug("No mainloop detected"); rc = EPROTO; } /* 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(rc != pcmk_ok) { crm_notice("Forcing immediate exit with status %d: %s", rc, pcmk_strerror(rc)); crm_write_blackbox(SIGTRAP, NULL); return crmd_fast_exit(rc); } /* 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); } clear_bit(fsa_input_register, R_MEMBERSHIP); g_list_free(fsa_message_queue); fsa_message_queue = NULL; free(pe_subsystem); pe_subsystem = NULL; free(te_subsystem); te_subsystem = NULL; free(cib_subsystem); cib_subsystem = NULL; if (metadata_hash) { crm_trace("Destroying reload cache with %d members", g_hash_table_size(metadata_hash)); g_hash_table_destroy(metadata_hash); metadata_hash = NULL; } election_fini(fsa_election); fsa_election = NULL; cib_delete(fsa_cib_conn); fsa_cib_conn = NULL; verify_stopped(fsa_state, LOG_WARNING); clear_bit(fsa_input_register, 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(stonith_reconnect); stonith_reconnect = NULL; mainloop_destroy_trigger(transition_trigger); transition_trigger = NULL; crm_client_cleanup(); crm_peer_destroy(); crm_timer_stop(transition_timer); crm_timer_stop(integration_timer); crm_timer_stop(finalization_timer); crm_timer_stop(election_trigger); election_timeout_stop(fsa_election); crm_timer_stop(shutdown_escalation_timer); crm_timer_stop(wait_timer); crm_timer_stop(recheck_timer); free(transition_timer); transition_timer = NULL; free(integration_timer); integration_timer = NULL; free(finalization_timer); finalization_timer = NULL; free(election_trigger); election_trigger = NULL; election_fini(fsa_election); free(shutdown_escalation_timer); shutdown_escalation_timer = NULL; free(wait_timer); wait_timer = NULL; free(recheck_timer); recheck_timer = NULL; 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(te_client_id); te_client_id = NULL; free(fsa_pe_ref); fsa_pe_ref = 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; crmd_drain_alerts(ctx); /* 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); #if SUPPORT_HEARTBEAT /* Do this only after g_main_loop_quit(). * * This interface was broken (incomplete) since it was introduced. * ->delete() does cleanup and free most of it, but it does not * actually remove and destroy the corresponding GSource, so the next * prepare/check iteratioin would find a corrupt (because partially * freed) GSource, and segfault. * * Apparently one was supposed to store the GSource as returned by * G_main_add_ll_cluster(), and g_source_destroy() that "by hand". * * But no-one ever did this, not even in the old hb code when this was * introduced. * * Note that fsa_cluster_conn was set as an "alias" to cluster->hb_conn * in do_ha_control() right after crm_cluster_connect(), and only * happens to still point at that object, because do_ha_control() does * not reset it to NULL after crm_cluster_disconnect() above does * reset cluster->hb_conn to NULL. * Not sure if that's something to cleanup, too. * * I'll try to fix this up in heartbeat proper, so ->delete * will actually remove, and destroy, and unref, and free this thing. * Doing so after g_main_loop_quit() is valid with both old, * and eventually fixed heartbeat. * * If we introduce the "by hand" destroy/remove/unref, * this may break again once heartbeat is fixed :-( * * -- Lars Ellenberg */ if (fsa_cluster_conn) { crm_trace("Deleting heartbeat api object"); fsa_cluster_conn->llc_ops->delete(fsa_cluster_conn); fsa_cluster_conn = NULL; } #endif /* Won't do anything yet, since we're inside it now */ g_main_loop_unref(mloop); crm_trace("Done %d", rc); } else { mainloop_destroy_signal(SIGCHLD); } /* Graceful */ return rc; } /* 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) { int exit_code = pcmk_ok; int log_level = LOG_INFO; const char *exit_type = "gracefully"; if (action & A_EXIT_1) { /* exit_code = pcmk_err_generic; */ log_level = LOG_ERR; exit_type = "forcefully"; exit_code = pcmk_err_generic; } verify_stopped(cur_state, LOG_ERR); do_crm_log(log_level, "Performing %s - %s exiting the CRMd", 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) { int was_error = 0; 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 and LRM objects"); fsa_cib_conn = cib_new(); lrm_state_init_local(); /* set up the timers */ transition_timer = calloc(1, sizeof(fsa_timer_t)); integration_timer = calloc(1, sizeof(fsa_timer_t)); finalization_timer = calloc(1, sizeof(fsa_timer_t)); election_trigger = calloc(1, sizeof(fsa_timer_t)); shutdown_escalation_timer = calloc(1, sizeof(fsa_timer_t)); wait_timer = calloc(1, sizeof(fsa_timer_t)); recheck_timer = calloc(1, sizeof(fsa_timer_t)); if (election_trigger != NULL) { election_trigger->source_id = 0; election_trigger->period_ms = -1; election_trigger->fsa_input = I_DC_TIMEOUT; election_trigger->callback = crm_timer_popped; election_trigger->repeat = FALSE; } else { was_error = TRUE; } if (transition_timer != NULL) { transition_timer->source_id = 0; transition_timer->period_ms = -1; transition_timer->fsa_input = I_PE_CALC; transition_timer->callback = crm_timer_popped; transition_timer->repeat = FALSE; } else { was_error = TRUE; } if (integration_timer != NULL) { integration_timer->source_id = 0; integration_timer->period_ms = -1; integration_timer->fsa_input = I_INTEGRATED; integration_timer->callback = crm_timer_popped; integration_timer->repeat = FALSE; } else { was_error = TRUE; } if (finalization_timer != NULL) { finalization_timer->source_id = 0; finalization_timer->period_ms = -1; finalization_timer->fsa_input = I_FINALIZED; finalization_timer->callback = crm_timer_popped; finalization_timer->repeat = FALSE; /* for possible enabling... a bug in the join protocol left * a slave in S_PENDING while we think its in S_NOT_DC * * raising I_FINALIZED put us into a transition loop which is * never resolved. * in this loop we continually send probes which the node * NACK's because its in S_PENDING * * if we have nodes where heartbeat is active but the * CRM is not... then this will be handled in the * integration phase */ finalization_timer->fsa_input = I_ELECTION; } else { was_error = TRUE; } if (shutdown_escalation_timer != NULL) { shutdown_escalation_timer->source_id = 0; shutdown_escalation_timer->period_ms = -1; shutdown_escalation_timer->fsa_input = I_STOP; shutdown_escalation_timer->callback = crm_timer_popped; shutdown_escalation_timer->repeat = FALSE; } else { was_error = TRUE; } if (wait_timer != NULL) { wait_timer->source_id = 0; wait_timer->period_ms = 2000; wait_timer->fsa_input = I_NULL; wait_timer->callback = crm_timer_popped; wait_timer->repeat = FALSE; } else { was_error = TRUE; } if (recheck_timer != NULL) { recheck_timer->source_id = 0; recheck_timer->period_ms = -1; recheck_timer->fsa_input = I_PE_CALC; recheck_timer->callback = crm_timer_popped; recheck_timer->repeat = FALSE; } else { was_error = TRUE; } /* set up the sub systems */ cib_subsystem = calloc(1, sizeof(struct crm_subsystem_s)); te_subsystem = calloc(1, sizeof(struct crm_subsystem_s)); pe_subsystem = calloc(1, sizeof(struct crm_subsystem_s)); if (cib_subsystem != NULL) { cib_subsystem->pid = -1; cib_subsystem->name = CRM_SYSTEM_CIB; cib_subsystem->flag_connected = R_CIB_CONNECTED; cib_subsystem->flag_required = R_CIB_REQUIRED; } else { was_error = TRUE; } if (te_subsystem != NULL) { te_subsystem->pid = -1; te_subsystem->name = CRM_SYSTEM_TENGINE; te_subsystem->flag_connected = R_TE_CONNECTED; te_subsystem->flag_required = R_TE_REQUIRED; } else { was_error = TRUE; } if (pe_subsystem != NULL) { pe_subsystem->pid = -1; pe_subsystem->path = CRM_DAEMON_DIR; pe_subsystem->name = CRM_SYSTEM_PENGINE; pe_subsystem->command = CRM_DAEMON_DIR "/" CRM_SYSTEM_PENGINE; pe_subsystem->args = NULL; pe_subsystem->flag_connected = R_PE_CONNECTED; pe_subsystem->flag_required = R_PE_REQUIRED; } else { was_error = TRUE; } if (was_error == FALSE && need_spawn_pengine_from_crmd()) { if (start_subsystem(pe_subsystem) == FALSE) { was_error = TRUE; } } if (was_error) { register_fsa_error(C_FSA_INTERNAL, I_ERROR, NULL); } } static int32_t crmd_ipc_accept(qb_ipcs_connection_t * c, uid_t uid, gid_t gid) { crm_trace("Connection %p", c); if (crm_client_new(c, uid, gid) == NULL) { return -EIO; } return 0; } static void crmd_ipc_created(qb_ipcs_connection_t * c) { crm_trace("Connection %p", c); } static int32_t crmd_ipc_dispatch(qb_ipcs_connection_t * c, void *data, size_t size) { uint32_t id = 0; uint32_t flags = 0; crm_client_t *client = crm_client_get(c); xmlNode *msg = crm_ipcs_recv(client, data, size, &id, &flags); crm_trace("Invoked: %s", crm_client_name(client)); crm_ipcs_send_ack(client, id, flags, "ack", __FUNCTION__, __LINE__); if (msg == NULL) { return 0; } #if ENABLE_ACL CRM_ASSERT(client->user != NULL); crm_acl_get_set_user(msg, F_CRM_USER, client->user); #endif crm_trace("Processing msg from %s", crm_client_name(client)); crm_log_xml_trace(msg, "CRMd[inbound]"); crm_xml_add(msg, F_CRM_SYS_FROM, client->id); if (crmd_authorize_message(msg, client, NULL)) { route_message(C_IPC_MESSAGE, msg); } trigger_fsa(fsa_source); free_xml(msg); return 0; } static int32_t crmd_ipc_closed(qb_ipcs_connection_t * c) { crm_client_t *client = crm_client_get(c); struct crm_subsystem_s *the_subsystem = NULL; if (client == NULL) { return 0; } crm_trace("Connection %p", c); if (client->userdata == NULL) { crm_trace("Client hadn't registered with us yet"); } else if (strcasecmp(CRM_SYSTEM_PENGINE, client->userdata) == 0) { the_subsystem = pe_subsystem; } else if (strcasecmp(CRM_SYSTEM_TENGINE, client->userdata) == 0) { the_subsystem = te_subsystem; } else if (strcasecmp(CRM_SYSTEM_CIB, client->userdata) == 0) { the_subsystem = cib_subsystem; } if (the_subsystem != NULL) { the_subsystem->source = NULL; the_subsystem->client = NULL; crm_info("Received HUP from %s:[%d]", the_subsystem->name, the_subsystem->pid); } else { /* else that was a transient client */ crm_trace("Received HUP from transient client"); } crm_trace("Disconnecting client %s (%p)", crm_client_name(client), client); free(client->userdata); crm_client_destroy(client); trigger_fsa(fsa_source); 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 = crmd_ipc_accept, .connection_created = crmd_ipc_created, .msg_process = crmd_ipc_dispatch, .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, LRM not connected (%.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) { /* try reading from HA */ crm_info("Delaying start, No peer data (%.16llx)", R_PEER_DATA); #if SUPPORT_HEARTBEAT if (is_heartbeat_cluster()) { HA_Message *msg = NULL; crm_trace("Looking for a HA message"); msg = fsa_cluster_conn->llc_ops->readmsg(fsa_cluster_conn, 0); if (msg != NULL) { crm_trace("There was a HA message"); ha_msg_del(msg); } } #endif crmd_fsa_stall(TRUE); return; } crm_debug("Init server comms"); ipcs = crmd_ipc_server_init(&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); } if (stonith_reconnect == NULL) { int dummy; stonith_reconnect = mainloop_add_trigger(G_PRIORITY_LOW, te_connect_stonith, &dummy); } set_bit(fsa_input_register, R_ST_REQUIRED); mainloop_set_trigger(stonith_reconnect); crm_notice("The local CRM is operational"); clear_bit(fsa_input_register, 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) { set_bit(fsa_input_register, R_IN_RECOVERY); crm_warn("Fast-tracking shutdown in response to errors"); register_fsa_input(C_FSA_INTERNAL, I_TERMINATE, NULL); } /* *INDENT-OFF* */ pe_cluster_option crmd_opts[] = { /* name, old-name, validate, values, default, short description, long description */ { "dc-version", NULL, "string", NULL, "none", NULL, "Version of Pacemaker on the cluster's DC.", "Includes the hash which identifies the exact changeset it was built from. Used for diagnostic purposes." }, { "cluster-infrastructure", NULL, "string", NULL, "heartbeat", NULL, "The messaging stack on which Pacemaker is currently running.", "Used for informational and diagnostic purposes." }, { XML_CONFIG_ATTR_DC_DEADTIME, "dc_deadtime", "time", NULL, "20s", &check_time, "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." }, { XML_CONFIG_ATTR_RECHECK, "cluster_recheck_interval", "time", "Zero disables polling. Positive values are an interval in seconds (unless other SI units are specified. eg. 5min)", "15min", &check_timer, "Polling interval for time based changes to options, resource parameters and constraints.", "The Cluster is primarily event driven, however the configuration can have elements that change based on time." " To ensure these changes take effect, we can optionally poll the cluster's status for changes." }, #ifdef RHEL7_COMPAT /* These options were superseded by the alerts feature and now are just an * alternate interface to it. It was never released upstream, but was * released in RHEL 7, so we allow it to be enabled at compile-time by * defining RHEL7_COMPAT. */ { "notification-agent", NULL, "string", NULL, "/dev/null", &check_script, "Deprecated", "Use alert path in alerts section instead" }, { "notification-recipient", NULL, "string", NULL, "", NULL, "Deprecated", "Use recipient value in alerts section instead" }, #endif { "load-threshold", NULL, "percentage", NULL, "80%", &check_utilization, "The maximum amount of system resources that should be used by nodes in the cluster", "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", &check_number, "The maximum number of jobs that can be scheduled per node. Defaults to 2x cores"}, { XML_CONFIG_ATTR_ELECTION_FAIL, "election_timeout", "time", NULL, "2min", &check_timer, "*** Advanced Use Only ***.", "If need to adjust this value, it probably indicates the presence of a bug." }, { XML_CONFIG_ATTR_FORCE_QUIT, "shutdown_escalation", "time", NULL, "20min", &check_timer, "*** Advanced Use Only ***.", "If need to adjust this value, it probably indicates the presence of a bug." }, { "crmd-integration-timeout", NULL, "time", NULL, "3min", &check_timer, "*** Advanced Use Only ***.", "If need to adjust this value, it probably indicates the presence of a bug." }, { "crmd-finalization-timeout", NULL, "time", NULL, "30min", &check_timer, "*** Advanced Use Only ***.", "If you need to adjust this value, it probably indicates the presence of a bug." }, { "crmd-transition-delay", NULL, "time", NULL, "0s", &check_timer, "*** Advanced Use Only ***\n" "Enabling this option will slow down cluster recovery under all conditions", "Delay cluster recovery for the configured interval to allow for additional/related events to occur.\n" "Useful if your configuration is sensitive to the order in which ping updates arrive." }, { "stonith-watchdog-timeout", NULL, "time", NULL, NULL, &check_sbd_timeout, "How long to wait before we can assume nodes are safely down", NULL }, { "stonith-max-attempts",NULL,"integer",NULL,"10",&check_positive_number, "How many times stonith can fail before it will no longer be attempted on a target" }, { "no-quorum-policy", "no_quorum_policy", "enum", "stop, freeze, ignore, suicide", "stop", &check_quorum, NULL, NULL }, #if SUPPORT_PLUGIN { XML_ATTR_EXPECTED_VOTES, NULL, "integer", NULL, "2", &check_number, "The number of nodes expected to be in the cluster", "Used to calculate quorum in openais based clusters." }, #endif }; /* *INDENT-ON* */ void crmd_metadata(void) { config_metadata("CRM Daemon", "1.0", "CRM Daemon Options", "This is a fake resource that details the options that can be configured for the CRM Daemon.", crmd_opts, DIMOF(crmd_opts)); } static void verify_crmd_options(GHashTable * options) { verify_all_options(options, crmd_opts, DIMOF(crmd_opts)); } static const char * crmd_pref(GHashTable * options, const char *name) { return get_cluster_pref(options, crmd_opts, DIMOF(crmd_opts), name); } static void config_query_callback(xmlNode * msg, int call_id, int rc, xmlNode * output, void *user_data) { #ifdef RHEL7_COMPAT const char *script = NULL; #endif 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"); set_bit(fsa_input_register, 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 = g_hash_table_new_full(crm_str_hash, g_str_equal, g_hash_destroy_str, g_hash_destroy_str); unpack_instance_attributes(crmconfig, crmconfig, XML_CIB_TAG_PROPSET, NULL, config_hash, CIB_OPTIONS_FIRST, FALSE, now); verify_crmd_options(config_hash); #ifdef RHEL7_COMPAT script = crmd_pref(config_hash, "notification-agent"); value = crmd_pref(config_hash, "notification-recipient"); crmd_enable_alerts(script, value); #endif value = crmd_pref(config_hash, XML_CONFIG_ATTR_DC_DEADTIME); election_trigger->period_ms = crm_get_msec(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_load_target = strtof(value, NULL) / 100; + throttle_set_load_target(strtof(value, NULL) / 100.0); } value = crmd_pref(config_hash, "no-quorum-policy"); if (safe_str_eq(value, "suicide") && pcmk_locate_sbd()) { no_quorum_suicide_escalation = TRUE; } - + 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_get_msec(value); /* How long to declare an election over - even if not everyone voted */ crm_debug("Shutdown escalation occurs after: %dms", shutdown_escalation_timer->period_ms); value = crmd_pref(config_hash, XML_CONFIG_ATTR_ELECTION_FAIL); election_timeout_set_period(fsa_election, crm_get_msec(value)); value = crmd_pref(config_hash, XML_CONFIG_ATTR_RECHECK); recheck_timer->period_ms = crm_get_msec(value); crm_debug("Checking for expired actions every %dms", recheck_timer->period_ms); value = crmd_pref(config_hash, "crmd-transition-delay"); transition_timer->period_ms = crm_get_msec(value); value = crmd_pref(config_hash, "crmd-integration-timeout"); integration_timer->period_ms = crm_get_msec(value); value = crmd_pref(config_hash, "crmd-finalization-timeout"); finalization_timer->period_ms = crm_get_msec(value); #if SUPPORT_COROSYNC if (is_classic_ais_cluster()) { value = crmd_pref(config_hash, XML_ATTR_EXPECTED_VOTES); crm_debug("Sending expected-votes=%s to corosync", value); send_cluster_text(crm_class_quorum, value, TRUE, NULL, crm_msg_ais); } #endif 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); parse_alerts(alerts); set_bit(fsa_input_register, 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_is_running(crmd_mainloop)) { if (is_set(fsa_input_register, R_SHUTDOWN)) { crm_err("Escalating the shutdown"); register_fsa_input_before(C_SHUTDOWN, I_ERROR, NULL); } else { set_bit(fsa_input_register, R_SHUTDOWN); register_fsa_input(C_SHUTDOWN, I_SHUTDOWN, NULL); if (shutdown_escalation_timer->period_ms < 1) { const char *value = crmd_pref(NULL, XML_CONFIG_ATTR_FORCE_QUIT); int msec = crm_get_msec(value); crm_debug("Using default shutdown escalation: %dms", msec); shutdown_escalation_timer->period_ms = msec; } /* can't rely on this... */ crm_notice("Shutting down cluster resource manager " CRM_XS " limit=%dms", shutdown_escalation_timer->period_ms); crm_timer_start(shutdown_escalation_timer); } } else { crm_info("exit from shutdown"); crmd_exit(pcmk_ok); } } diff --git a/crmd/throttle.c b/crmd/throttle.c index ce330fe6d7..90ddb909ed 100644 --- a/crmd/throttle.c +++ b/crmd/throttle.c @@ -1,676 +1,570 @@ /* * Copyright (C) 2013 Andrew Beekhof * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public * License as published by the Free Software Foundation; either * version 2 of the License, or (at your option) any later version. * * This software is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * General Public License for more details. * * You should have received a copy of the GNU General Public * License along with this library; if not, write to the Free Software * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */ #include #include #include #include #include #include #include #include #include #include #include -enum throttle_state_e -{ +enum throttle_state_e { throttle_extreme = 0x1000, throttle_high = 0x0100, throttle_med = 0x0010, throttle_low = 0x0001, throttle_none = 0x0000, }; -struct throttle_record_s -{ - int max; - enum throttle_state_e mode; - char *node; +struct throttle_record_s { + int max; + enum throttle_state_e mode; + char *node; }; -int throttle_job_max = 0; -float throttle_load_target = 0.0; +static int throttle_job_max = 0; +static float throttle_load_target = 0.0; #define THROTTLE_FACTOR_LOW 1.2 #define THROTTLE_FACTOR_MEDIUM 1.6 #define THROTTLE_FACTOR_HIGH 2.0 -GHashTable *throttle_records = NULL; -mainloop_timer_t *throttle_timer = NULL; - -int throttle_num_cores(void) -{ - static int cores = 0; - char buffer[256]; - FILE *stream = NULL; - const char *cpufile = "/proc/cpuinfo"; - - if(cores) { - return cores; - } - stream = fopen(cpufile, "r"); - if(stream == NULL) { - int rc = errno; - crm_warn("Couldn't read %s, assuming a single processor: %s (%d)", cpufile, pcmk_strerror(rc), rc); - return 1; - } - - while (fgets(buffer, sizeof(buffer), stream)) { - if(strstr(buffer, "processor") == buffer) { - cores++; - } - } - - fclose(stream); - - if(cores == 0) { - crm_warn("No processors found in %s, assuming 1", cpufile); - return 1; - } - - return cores; -} +static GHashTable *throttle_records = NULL; +static mainloop_timer_t *throttle_timer = NULL; /*! * \internal * \brief Return name of /proc file containing the CIB deamon's load statistics * * \return Newly allocated memory with file name on success, NULL otherwise * * \note It is the caller's responsibility to free the return value. * This will return NULL if the daemon is being run via valgrind. * This should be called only on Linux systems. */ -static char *find_cib_loadfile(void) +static char * +find_cib_loadfile(void) { int pid = crm_procfs_pid_of("cib"); return pid? crm_strdup_printf("/proc/%d/stat", pid) : NULL; } -static bool throttle_cib_load(float *load) +static bool +throttle_cib_load(float *load) { /* /proc/[pid]/stat Status information about the process. This is used by ps(1). It is defined in /usr/src/linux/fs/proc/array.c. The fields, in order, with their proper scanf(3) format specifiers, are: pid %d (1) The process ID. comm %s (2) The filename of the executable, in parentheses. This is visible whether or not the executable is swapped out. state %c (3) One character from the string "RSDZTW" where R is running, S is sleeping in an interruptible wait, D is waiting in uninterruptible disk sleep, Z is zombie, T is traced or stopped (on a signal), and W is paging. ppid %d (4) The PID of the parent. pgrp %d (5) The process group ID of the process. session %d (6) The session ID of the process. tty_nr %d (7) The controlling terminal of the process. (The minor device number is contained in the combination of bits 31 to 20 and 7 to 0; the major device number is in bits 15 to 8.) tpgid %d (8) The ID of the foreground process group of the controlling terminal of the process. flags %u (%lu before Linux 2.6.22) (9) The kernel flags word of the process. For bit meanings, see the PF_* defines in the Linux kernel source file include/linux/sched.h. Details depend on the kernel version. minflt %lu (10) The number of minor faults the process has made which have not required loading a memory page from disk. cminflt %lu (11) The number of minor faults that the process's waited-for children have made. majflt %lu (12) The number of major faults the process has made which have required loading a memory page from disk. cmajflt %lu (13) The number of major faults that the process's waited-for children have made. utime %lu (14) Amount of time that this process has been scheduled in user mode, measured in clock ticks (divide by sysconf(_SC_CLK_TCK)). This includes guest time, guest_time (time spent running a virtual CPU, see below), so that applications that are not aware of the guest time field do not lose that time from their calculations. stime %lu (15) Amount of time that this process has been scheduled in kernel mode, measured in clock ticks (divide by sysconf(_SC_CLK_TCK)). */ static char *loadfile = NULL; static time_t last_call = 0; static long ticks_per_s = 0; static unsigned long last_utime, last_stime; char buffer[64*1024]; FILE *stream = NULL; time_t now = time(NULL); if(load == NULL) { return FALSE; } else { *load = 0.0; } if(loadfile == NULL) { last_call = 0; last_utime = 0; last_stime = 0; loadfile = find_cib_loadfile(); if (loadfile == NULL) { crm_warn("Couldn't find CIB load file"); return FALSE; } ticks_per_s = sysconf(_SC_CLK_TCK); crm_trace("Found %s", loadfile); } stream = fopen(loadfile, "r"); if(stream == NULL) { int rc = errno; crm_warn("Couldn't read %s: %s (%d)", loadfile, pcmk_strerror(rc), rc); free(loadfile); loadfile = NULL; return FALSE; } if(fgets(buffer, sizeof(buffer), stream)) { char *comm = calloc(1, 256); char state = 0; int rc = 0, pid = 0, ppid = 0, pgrp = 0, session = 0, tty_nr = 0, tpgid = 0; unsigned long flags = 0, minflt = 0, cminflt = 0, majflt = 0, cmajflt = 0, utime = 0, stime = 0; rc = sscanf(buffer, "%d %[^ ] %c %d %d %d %d %d %lu %lu %lu %lu %lu %lu %lu", &pid, comm, &state, &ppid, &pgrp, &session, &tty_nr, &tpgid, &flags, &minflt, &cminflt, &majflt, &cmajflt, &utime, &stime); free(comm); if(rc != 15) { crm_err("Only %d of 15 fields found in %s", rc, loadfile); fclose(stream); return FALSE; } else if(last_call > 0 && last_call < now && last_utime <= utime && last_stime <= stime) { time_t elapsed = now - last_call; unsigned long delta_utime = utime - last_utime; unsigned long delta_stime = stime - last_stime; *load = (delta_utime + delta_stime); /* Cast to a float before division */ *load /= ticks_per_s; *load /= elapsed; crm_debug("cib load: %f (%lu ticks in %lds)", *load, delta_utime + delta_stime, (long)elapsed); } else { crm_debug("Init %lu + %lu ticks at %ld (%lu tps)", utime, stime, (long)now, ticks_per_s); } last_call = now; last_utime = utime; last_stime = stime; fclose(stream); return TRUE; } fclose(stream); return FALSE; } -static bool throttle_load_avg(float *load) +static bool +throttle_load_avg(float *load) { char buffer[256]; FILE *stream = NULL; const char *loadfile = "/proc/loadavg"; if(load == NULL) { return FALSE; } stream = fopen(loadfile, "r"); if(stream == NULL) { int rc = errno; crm_warn("Couldn't read %s: %s (%d)", loadfile, pcmk_strerror(rc), rc); return FALSE; } if(fgets(buffer, sizeof(buffer), stream)) { char *nl = strstr(buffer, "\n"); /* Grab the 1-minute average, ignore the rest */ *load = strtof(buffer, NULL); if(nl) { nl[0] = 0; } - crm_debug("Current load is %f (full: %s)", *load, buffer); fclose(stream); return TRUE; } fclose(stream); return FALSE; } -static bool throttle_io_load(float *load, unsigned int *blocked) +/*! + * \internal + * \brief Check a load value against throttling thresholds + * + * \param[in] load Load value to check + * \param[in] desc Description of metric (for logging) + * \param[in] thresholds Low/medium/high/extreme thresholds + * + * \return Throttle mode corresponding to load value + */ +static enum throttle_state_e +throttle_check_thresholds(float load, const char *desc, float thresholds[4]) { - char buffer[64*1024]; - FILE *stream = NULL; - const char *loadfile = "/proc/stat"; - - if(load == NULL) { - return FALSE; - } - - stream = fopen(loadfile, "r"); - if(stream == NULL) { - int rc = errno; - crm_warn("Couldn't read %s: %s (%d)", loadfile, pcmk_strerror(rc), rc); - return FALSE; - } - - if(fgets(buffer, sizeof(buffer), stream)) { - /* Borrowed from procps-ng's sysinfo.c */ - - char *b = NULL; - unsigned long long cpu_use = 0; - unsigned long long cpu_nic = 0; - unsigned long long cpu_sys = 0; - unsigned long long cpu_idl = 0; - unsigned long long cpu_iow = 0; /* not separated out until the 2.5.41 kernel */ - unsigned long long cpu_xxx = 0; /* not separated out until the 2.6.0-test4 kernel */ - unsigned long long cpu_yyy = 0; /* not separated out until the 2.6.0-test4 kernel */ - unsigned long long cpu_zzz = 0; /* not separated out until the 2.6.11 kernel */ - - long long divo2 = 0; - long long duse = 0; - long long dsys = 0; - long long didl =0; - long long diow =0; - long long dstl = 0; - long long Div = 0; - - b = strstr(buffer, "cpu "); - if(b) sscanf(b, "cpu %Lu %Lu %Lu %Lu %Lu %Lu %Lu %Lu", - &cpu_use, &cpu_nic, &cpu_sys, &cpu_idl, &cpu_iow, &cpu_xxx, &cpu_yyy, &cpu_zzz); - - if(blocked) { - b = strstr(buffer, "procs_blocked "); - if(b) sscanf(b, "procs_blocked %u", blocked); - } + if (load > thresholds[3]) { + crm_notice("Extreme %s detected: %f", desc, load); + return throttle_extreme; - duse = cpu_use + cpu_nic; - dsys = cpu_sys + cpu_xxx + cpu_yyy; - didl = cpu_idl; - diow = cpu_iow; - dstl = cpu_zzz; - Div = duse + dsys + didl + diow + dstl; - if (!Div) Div = 1, didl = 1; - divo2 = Div / 2UL; + } else if (load > thresholds[2]) { + crm_notice("High %s detected: %f", desc, load); + return throttle_high; - /* vmstat output: - * - * procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu---- - * r b swpd free buff cache si so bi bo in cs us sy id wa - * 1 0 5537800 958592 204180 1737740 1 1 12 15 0 0 2 1 97 0 - * - * The last four columns are calculated as: - * - * (unsigned)( (100*duse + divo2) / Div ), - * (unsigned)( (100*dsys + divo2) / Div ), - * (unsigned)( (100*didl + divo2) / Div ), - * (unsigned)( (100*diow + divo2) / Div ) - * - */ - *load = (diow + divo2) / Div; - crm_debug("Current IO load is %f", *load); + } else if (load > thresholds[1]) { + crm_info("Moderate %s detected: %f", desc, load); + return throttle_med; - fclose(stream); - return TRUE; + } else if (load > thresholds[0]) { + crm_debug("Noticeable %s detected: %f", desc, load); + return throttle_low; } - fclose(stream); - return FALSE; + crm_trace("Negligible %s detected: %f", desc, load); + return throttle_none; } static enum throttle_state_e throttle_handle_load(float load, const char *desc, int cores) { - float adjusted_load = load; - - if(cores <= 0) { - /* No fudging of the supplied load value */ + float normalize; + float thresholds[4]; - } else if(cores == 1) { + if (cores == 1) { /* On a single core machine, a load of 1.0 is already too high */ - adjusted_load = load * THROTTLE_FACTOR_MEDIUM; + normalize = 0.6; } else { /* Normalize the load to be per-core */ - adjusted_load = load / cores; + normalize = cores; } + thresholds[0] = throttle_load_target * normalize * THROTTLE_FACTOR_LOW; + thresholds[1] = throttle_load_target * normalize * THROTTLE_FACTOR_MEDIUM; + thresholds[2] = throttle_load_target * normalize * THROTTLE_FACTOR_HIGH; + thresholds[3] = load + 1.0; /* never extreme */ - if(adjusted_load > THROTTLE_FACTOR_HIGH * throttle_load_target) { - crm_notice("High %s detected: %f", desc, load); - return throttle_high; - - } else if(adjusted_load > THROTTLE_FACTOR_MEDIUM * throttle_load_target) { - crm_info("Moderate %s detected: %f", desc, load); - return throttle_med; - - } else if(adjusted_load > THROTTLE_FACTOR_LOW * throttle_load_target) { - crm_debug("Noticeable %s detected: %f", desc, load); - return throttle_low; - } - - crm_trace("Negligible %s detected: %f", desc, adjusted_load); - return throttle_none; + return throttle_check_thresholds(load, desc, thresholds); } static enum throttle_state_e throttle_mode(void) { - int cores; + unsigned int cores; float load; - unsigned int blocked = 0; + float thresholds[4]; enum throttle_state_e mode = throttle_none; #if defined(ON_BSD) || defined(ON_SOLARIS) return throttle_none; #endif - cores = throttle_num_cores(); + cores = crm_procfs_num_cores(); if(throttle_cib_load(&load)) { float cib_max_cpu = 0.95; - const char *desc = "CIB load"; - /* The CIB is a single threaded task and thus cannot consume + + /* The CIB is a single-threaded task and thus cannot consume * more than 100% of a CPU (and 1/cores of the overall system * load). * - * On a many cored system, the CIB might therefor be maxed out + * On a many-cored system, the CIB might therefore be maxed out * (causing operations to fail or appear to fail) even though * the overall system load is still reasonable. * - * Therefor the 'normal' thresholds can not apply here and we + * Therefore, the 'normal' thresholds can not apply here, and we * need a special case. */ if(cores == 1) { cib_max_cpu = 0.4; } if(throttle_load_target > 0.0 && throttle_load_target < cib_max_cpu) { cib_max_cpu = throttle_load_target; } - if(load > 1.5 * cib_max_cpu) { - /* Can only happen on machines with a low number of cores */ - crm_notice("Extreme %s detected: %f", desc, load); - mode |= throttle_extreme; - - } else if(load > cib_max_cpu) { - crm_notice("High %s detected: %f", desc, load); - mode |= throttle_high; - - } else if(load > cib_max_cpu * 0.9) { - crm_info("Moderate %s detected: %f", desc, load); - mode |= throttle_med; + thresholds[0] = cib_max_cpu * 0.8; + thresholds[1] = cib_max_cpu * 0.9; + thresholds[2] = cib_max_cpu; + /* Can only happen on machines with a low number of cores */ + thresholds[3] = cib_max_cpu * 1.5; - } else if(load > cib_max_cpu * 0.8) { - crm_debug("Noticeable %s detected: %f", desc, load); - mode |= throttle_low; - - } else { - crm_trace("Negligible %s detected: %f", desc, load); - } + mode |= throttle_check_thresholds(load, "CIB load", thresholds); } if(throttle_load_target <= 0) { /* If we ever make this a valid value, the cluster will at least behave as expected */ return mode; } if(throttle_load_avg(&load)) { + crm_debug("Current load is %f across %u core(s)", load, cores); mode |= throttle_handle_load(load, "CPU load", cores); } - if(throttle_io_load(&load, &blocked)) { - mode |= throttle_handle_load(load, "IO load", 0); - mode |= throttle_handle_load(blocked, "blocked IO ratio", cores); - } - if(mode & throttle_extreme) { return throttle_extreme; } else if(mode & throttle_high) { return throttle_high; } else if(mode & throttle_med) { return throttle_med; } else if(mode & throttle_low) { return throttle_low; } return throttle_none; } static void throttle_send_command(enum throttle_state_e mode) { xmlNode *xml = NULL; static enum throttle_state_e last = -1; if(mode != last) { crm_info("New throttle mode: %.4x (was %.4x)", mode, last); last = mode; xml = create_request(CRM_OP_THROTTLE, NULL, NULL, CRM_SYSTEM_CRMD, CRM_SYSTEM_CRMD, NULL); crm_xml_add_int(xml, F_CRM_THROTTLE_MODE, mode); crm_xml_add_int(xml, F_CRM_THROTTLE_MAX, throttle_job_max); send_cluster_message(NULL, crm_msg_crmd, xml, TRUE); free_xml(xml); } } static gboolean throttle_timer_cb(gpointer data) { static bool send_updates = FALSE; enum throttle_state_e now = throttle_none; if(send_updates) { now = throttle_mode(); throttle_send_command(now); } else if(compare_version(fsa_our_dc_version, "3.0.8") < 0) { /* Optimize for the true case */ crm_trace("DC version %s doesn't support throttling", fsa_our_dc_version); } else { send_updates = TRUE; now = throttle_mode(); throttle_send_command(now); } return TRUE; } static void throttle_record_free(gpointer p) { struct throttle_record_s *r = p; free(r->node); free(r); } void -throttle_update_job_max(const char *preference) +throttle_set_load_target(float target) +{ + throttle_load_target = target; +} + +void +throttle_update_job_max(const char *preference) { int max = 0; - throttle_job_max = 2 * throttle_num_cores(); + throttle_job_max = 2 * crm_procfs_num_cores(); if(preference) { /* Global preference from the CIB */ max = crm_int_helper(preference, NULL); if(max > 0) { throttle_job_max = max; } } preference = getenv("LRMD_MAX_CHILDREN"); if(preference) { /* Legacy env variable */ max = crm_int_helper(preference, NULL); if(max > 0) { throttle_job_max = max; } } preference = getenv("PCMK_node_action_limit"); if(preference) { /* Per-node override */ max = crm_int_helper(preference, NULL); if(max > 0) { throttle_job_max = max; } } } - void throttle_init(void) { if(throttle_records == NULL) { throttle_records = g_hash_table_new_full( crm_str_hash, g_str_equal, NULL, throttle_record_free); throttle_timer = mainloop_timer_add("throttle", 30 * 1000, TRUE, throttle_timer_cb, NULL); } throttle_update_job_max(NULL); mainloop_timer_start(throttle_timer); } void throttle_fini(void) { mainloop_timer_del(throttle_timer); throttle_timer = NULL; g_hash_table_destroy(throttle_records); throttle_records = NULL; } - int throttle_get_total_job_limit(int l) { /* Cluster-wide limit */ GHashTableIter iter; int limit = l; int peers = crm_active_peers(); struct throttle_record_s *r = NULL; g_hash_table_iter_init(&iter, throttle_records); while (g_hash_table_iter_next(&iter, NULL, (gpointer *) &r)) { switch(r->mode) { case throttle_extreme: if(limit == 0 || limit > peers/4) { limit = QB_MAX(1, peers/4); } break; case throttle_high: if(limit == 0 || limit > peers/2) { limit = QB_MAX(1, peers/2); } break; default: break; } } if(limit == l) { /* crm_trace("No change to batch-limit=%d", limit); */ } else if(l == 0) { crm_trace("Using batch-limit=%d", limit); } else { crm_trace("Using batch-limit=%d instead of %d", limit, l); } return limit; } int throttle_get_job_limit(const char *node) { int jobs = 1; struct throttle_record_s *r = NULL; r = g_hash_table_lookup(throttle_records, node); if(r == NULL) { r = calloc(1, sizeof(struct throttle_record_s)); r->node = strdup(node); r->mode = throttle_low; r->max = throttle_job_max; crm_trace("Defaulting to local values for unknown node %s", node); g_hash_table_insert(throttle_records, r->node, r); } switch(r->mode) { case throttle_extreme: case throttle_high: jobs = 1; /* At least one job must always be allowed */ break; case throttle_med: jobs = QB_MAX(1, r->max / 4); break; case throttle_low: jobs = QB_MAX(1, r->max / 2); break; case throttle_none: jobs = QB_MAX(1, r->max); break; default: crm_err("Unknown throttle mode %.4x on %s", r->mode, node); break; } return jobs; } void throttle_update(xmlNode *xml) { int max = 0; enum throttle_state_e mode = 0; struct throttle_record_s *r = NULL; const char *from = crm_element_value(xml, F_CRM_HOST_FROM); crm_element_value_int(xml, F_CRM_THROTTLE_MODE, (int*)&mode); crm_element_value_int(xml, F_CRM_THROTTLE_MAX, &max); r = g_hash_table_lookup(throttle_records, from); if(r == NULL) { r = calloc(1, sizeof(struct throttle_record_s)); r->node = strdup(from); g_hash_table_insert(throttle_records, r->node, r); } r->max = max; r->mode = mode; crm_debug("Host %s supports a maximum of %d jobs and throttle mode %.4x. New job limit is %d", from, max, mode, throttle_get_job_limit(from)); } - diff --git a/crmd/throttle.h b/crmd/throttle.h index bdb33e75bb..9d1b97b89a 100644 --- a/crmd/throttle.h +++ b/crmd/throttle.h @@ -1,29 +1,27 @@ /* * Copyright (C) 2013 Andrew Beekhof * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2 of the License, or (at your option) any later version. * * This software is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * General Public License for more details. * * You should have received a copy of the GNU Lesser General Public * License along with this library; if not, write to the Free Software * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */ -extern float throttle_load_target; - void throttle_init(void); void throttle_fini(void); -int throttle_num_cores(void); +void throttle_set_load_target(float target); void throttle_update(xmlNode *xml); void throttle_update_job_max(const char *preference); int throttle_get_job_limit(const char *node); int throttle_get_total_job_limit(int l); diff --git a/include/crm/common/internal.h b/include/crm/common/internal.h index 475128f2f4..c34b03b890 100644 --- a/include/crm/common/internal.h +++ b/include/crm/common/internal.h @@ -1,123 +1,124 @@ /* * Copyright (C) 2015 * Andrew Beekhof * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU Lesser General Public License as published by * the Free Software Foundation; either version 2 of the License, or * (at your option) any later version. * * This program is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public * License along with this library; if not, write to the Free Software * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */ /*! * \file * \brief internal common utilities * \ingroup core * \note Public APIs are declared in util.h */ #ifndef CRM_COMMON_INTERNAL__H #define CRM_COMMON_INTERNAL__H #include /* for gboolean */ #include /* for struct dirent */ #include /* for uid_t and gid_t */ #include /* internal I/O utilities (from io.c) */ char *generate_series_filename(const char *directory, const char *series, int sequence, gboolean bzip); int get_last_sequence(const char *directory, const char *series); void write_last_sequence(const char *directory, const char *series, int sequence, int max); int crm_chown_last_sequence(const char *directory, const char *series, uid_t uid, gid_t gid); gboolean crm_is_writable(const char *dir, const char *file, const char *user, const char *group, gboolean need_both); void crm_sync_directory(const char *name); char *crm_read_contents(const char *filename); int crm_write_sync(int fd, const char *contents); /* internal procfs utilities (from procfs.c) */ int crm_procfs_process_info(struct dirent *entry, char *name, int *pid); int crm_procfs_pid_of(const char *name); +unsigned int crm_procfs_num_cores(void); /* internal XML schema functions (from xml.c) */ void crm_schema_init(void); void crm_schema_cleanup(void); /* internal generic string functions (from strings.c) */ char *crm_concat(const char *prefix, const char *suffix, char join); void g_hash_destroy_str(gpointer data); long long crm_int_helper(const char *text, char **end_text); gboolean crm_ends_with(const char *s, const char *match); char *add_list_element(char *list, const char *value); bool crm_compress_string(const char *data, int length, int max, char **result, unsigned int *result_len); static inline int crm_strlen_zero(const char *s) { return !s || *s == '\0'; } /* convenience functions for failure-related node attributes */ #define CRM_FAIL_COUNT_PREFIX "fail-count" #define CRM_LAST_FAILURE_PREFIX "last-failure" /*! * \internal * \brief Generate a failure-related node attribute name for a resource * * \param[in] prefix Start of attribute name * \param[in] rsc_id Resource name * \param[in] op Operation name * \param[in] interval Operation interval * * \return Newly allocated string with attribute name * * \note Failure attributes are named like PREFIX-RSC#OP_INTERVAL (for example, * "fail-count-myrsc#monitor_30000"). The '#' is used because it is not * a valid character in a resource ID, to reliably distinguish where the * operation name begins. The '_' is used simply to be more comparable to * action labels like "myrsc_monitor_30000". */ static inline char * crm_fail_attr_name(const char *prefix, const char *rsc_id, const char *op, int interval) { CRM_CHECK(prefix && rsc_id && op, return NULL); return crm_strdup_printf("%s-%s#%s_%d", prefix, rsc_id, op, interval); } static inline char * crm_failcount_name(const char *rsc_id, const char *op, int interval) { return crm_fail_attr_name(CRM_FAIL_COUNT_PREFIX, rsc_id, op, interval); } static inline char * crm_lastfailure_name(const char *rsc_id, const char *op, int interval) { return crm_fail_attr_name(CRM_LAST_FAILURE_PREFIX, rsc_id, op, interval); } #endif /* CRM_COMMON_INTERNAL__H */ diff --git a/lib/common/procfs.c b/lib/common/procfs.c index 12d01ff7a7..fbbf9ebdd1 100644 --- a/lib/common/procfs.c +++ b/lib/common/procfs.c @@ -1,142 +1,172 @@ /* * Copyright (C) 2015 Andrew Beekhof * * This library is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * * This library is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public * License along with this library; if not, write to the Free Software * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */ #include #ifndef _GNU_SOURCE # define _GNU_SOURCE #endif #include #include #include #include #include #include +#include /*! * \internal * \brief Get process ID and name associated with a /proc directory entry * * \param[in] entry Directory entry (must be result of readdir() on /proc) * \param[out] name If not NULL, a char[64] to hold the process name * \param[out] pid If not NULL, will be set to process ID of entry * * \return 0 on success, -1 if entry is not for a process or info not found * * \note This should be called only on Linux systems, as not all systems that * support /proc store process names and IDs in the same way. */ int crm_procfs_process_info(struct dirent *entry, char *name, int *pid) { int fd, local_pid; FILE *file; struct stat statbuf; char key[16] = { 0 }, procpath[128] = { 0 }; /* We're only interested in entries whose name is a PID, * so skip anything non-numeric or that is too long. * * 114 = 128 - strlen("/proc/") - strlen("/status") - 1 */ local_pid = atoi(entry->d_name); if ((local_pid <= 0) || (strlen(entry->d_name) > 114)) { return -1; } if (pid) { *pid = local_pid; } /* Get this entry's file information */ strcpy(procpath, "/proc/"); strcat(procpath, entry->d_name); fd = open(procpath, O_RDONLY); if (fd < 0 ) { return -1; } if (fstat(fd, &statbuf) < 0) { close(fd); return -1; } close(fd); /* We're only interested in subdirectories */ if (!S_ISDIR(statbuf.st_mode)) { return -1; } /* Read the first entry ("Name:") from the process's status file. * We could handle the valgrind case if we parsed the cmdline file * instead, but that's more of a pain than it's worth. */ if (name != NULL) { strcat(procpath, "/status"); file = fopen(procpath, "r"); if (!file) { return -1; } if ((fscanf(file, "%15s%63s", key, name) != 2) || safe_str_neq(key, "Name:")) { fclose(file); return -1; } fclose(file); } return 0; } /*! * \internal * \brief Return process ID of a named process * * \param[in] name Process name (as used in /proc/.../status) * * \return Process ID of named process if running, 0 otherwise * * \note This will return 0 if the process is being run via valgrind. * This should be called only on Linux systems. */ int crm_procfs_pid_of(const char *name) { DIR *dp; struct dirent *entry; int pid = 0; char entry_name[64] = { 0 }; dp = opendir("/proc"); if (dp == NULL) { crm_notice("Can not read /proc directory to track existing components"); return 0; } while ((entry = readdir(dp)) != NULL) { if ((crm_procfs_process_info(entry, entry_name, &pid) == 0) && safe_str_eq(entry_name, name) && (crm_pid_active(pid, NULL) == 1)) { crm_info("Found %s active as process %d", name, pid); break; } pid = 0; } closedir(dp); return pid; } + +/*! + * \internal + * \brief Calculate number of logical CPU cores from procfs + * + * \return Number of cores (or 1 if unable to determine) + */ +unsigned int +crm_procfs_num_cores(void) +{ + int cores = 0; + FILE *stream = NULL; + + /* Parse /proc/stat instead of /proc/cpuinfo because it's smaller */ + stream = fopen("/proc/stat", "r"); + if (stream == NULL) { + crm_perror(LOG_INFO, "Could not open /proc/stat"); + } else { + char buffer[2048]; + + while (fgets(buffer, sizeof(buffer), stream)) { + if (!strncmp(buffer, "cpu", 3) && isdigit(buffer[3])) { + ++cores; + } + } + fclose(stream); + } + return cores? cores : 1; +}