/*****************************************************************************\ * slurm_acct_gather_energy.c - implementation-independent job energy * accounting plugin definitions ***************************************************************************** * Copyright (C) 2012-2019 SchedMD LLC * Copyright (C) 2012 Bull-HN-PHX. * Written by Bull-HN-PHX/d.rusak, Danny Auble * * This file is part of Slurm, a resource management program. * For details, see . * Please also read the included file: DISCLAIMER. * * Slurm 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. * * In addition, as a special exception, the copyright holders give permission * to link the code of portions of this program with the OpenSSL library under * certain conditions as described in each individual source file, and * distribute linked combinations including the two. You must obey the GNU * General Public License in all respects for all of the code used other than * OpenSSL. If you modify file(s) with this exception, you may extend this * exception to your version of the file(s), but you are not obligated to do * so. If you do not wish to do so, delete this exception statement from your * version. If you delete this exception statement from all source files in * the program, then also delete it here. * * Slurm 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 Slurm; if not, write to the Free Software Foundation, Inc., * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. \*****************************************************************************/ #ifdef HAVE_CONFIG_H # include "config.h" #endif #include #include #include #if HAVE_SYS_PRCTL_H # include #endif #include "src/common/macros.h" #include "src/common/plugin.h" #include "src/common/plugrack.h" #include "src/common/slurm_protocol_api.h" #include "src/common/xmalloc.h" #include "src/common/xstring.h" #include "src/common/slurm_acct_gather_energy.h" #include "src/slurmd/slurmstepd/slurmstepd_job.h" /* ** Define slurm-specific aliases for use by plugins, see slurm_xlator.h ** for details. */ strong_alias(acct_gather_energy_destroy, slurm_acct_gather_energy_destroy); typedef struct slurm_acct_gather_energy_ops { int (*update_node_energy) (void); int (*get_data) (enum acct_energy_type data_type, void *data); int (*set_data) (enum acct_energy_type data_type, void *data); void (*conf_options) (s_p_options_t **full_options, int *full_options_cnt); void (*conf_set) (int context_id_in, s_p_hashtbl_t *tbl); void (*conf_values) (List *data); } slurm_acct_gather_energy_ops_t; /* * These strings must be kept in the same order as the fields * declared for slurm_acct_gather_energy_ops_t. */ static const char *syms[] = { "acct_gather_energy_p_update_node_energy", "acct_gather_energy_p_get_data", "acct_gather_energy_p_set_data", "acct_gather_energy_p_conf_options", "acct_gather_energy_p_conf_set", "acct_gather_energy_p_conf_values", }; static slurm_acct_gather_energy_ops_t *ops; static plugin_context_t **g_context = NULL; static int g_context_num = -1; static pthread_mutex_t g_context_lock = PTHREAD_MUTEX_INITIALIZER; static bool init_run = false; static bool acct_shutdown = true; static int freq = 0; static pthread_t watch_node_thread_id = 0; static acct_gather_profile_timer_t *profile_timer = &acct_gather_profile_timer[PROFILE_ENERGY]; static void *_watch_node(void *arg) { int delta = profile_timer->freq - 1; #if HAVE_SYS_PRCTL_H if (prctl(PR_SET_NAME, "acctg_energy", NULL, NULL, NULL) < 0) { error("%s: cannot set my name to %s %m", __func__, "acctg_energy"); } #endif while (init_run && acct_gather_profile_test()) { /* Do this until shutdown is requested */ slurm_mutex_lock(&g_context_lock); for (int i = 0; i < g_context_num; i++) { if (!g_context[i]) continue; (*(ops[i].set_data))(ENERGY_DATA_PROFILE, &delta); } slurm_mutex_unlock(&g_context_lock); slurm_mutex_lock(&profile_timer->notify_mutex); slurm_cond_wait(&profile_timer->notify, &profile_timer->notify_mutex); slurm_mutex_unlock(&profile_timer->notify_mutex); } return NULL; } extern int slurm_acct_gather_energy_init(void) { int retval = SLURM_SUCCESS; char *plugin_type = "acct_gather_energy"; char *full_plugin_type = NULL; char *last = NULL, *plugin_entry, *type = NULL; if (init_run && (g_context_num >= 0)) return retval; slurm_mutex_lock(&g_context_lock); if (g_context_num >= 0) goto done; full_plugin_type = slurm_get_acct_gather_energy_type(); g_context_num = 0; /* mark it before anything else */ plugin_entry = full_plugin_type; while ((type = strtok_r(plugin_entry, ",", &last))) { xrealloc(ops, sizeof(slurm_acct_gather_energy_ops_t) * (g_context_num + 1)); xrealloc(g_context, (sizeof(plugin_context_t *) * (g_context_num + 1))); if (!xstrncmp(type, "acct_gather_energy/", 19)) type += 19; /* backward compatibility */ type = xstrdup_printf("%s/%s", plugin_type, type); g_context[g_context_num] = plugin_context_create( plugin_type, type, (void **)&ops[g_context_num], syms, sizeof(syms)); if (!g_context[g_context_num]) { error("cannot create %s context for %s", plugin_type, type); xfree(type); retval = SLURM_ERROR; break; } xfree(type); g_context_num++; plugin_entry = NULL; /* for next iteration */ } xfree(full_plugin_type); init_run = true; done: slurm_mutex_unlock(&g_context_lock); if (retval == SLURM_SUCCESS) retval = acct_gather_conf_init(); if (retval != SLURM_SUCCESS) fatal("can not open the %s plugin", type); xfree(type); return retval; } extern int acct_gather_energy_fini(void) { int rc2, rc = SLURM_SUCCESS; slurm_mutex_lock(&g_context_lock); init_run = false; if (watch_node_thread_id) { slurm_mutex_unlock(&g_context_lock); slurm_mutex_lock(&profile_timer->notify_mutex); slurm_cond_signal(&profile_timer->notify); slurm_mutex_unlock(&profile_timer->notify_mutex); pthread_join(watch_node_thread_id, NULL); slurm_mutex_lock(&g_context_lock); } for (int i = 0; i < g_context_num; i++) { if (!g_context[i]) continue; rc2 = plugin_context_destroy(g_context[i]); if (rc2 != SLURM_SUCCESS) { debug("%s: %s: %s", __func__, g_context[i]->type, slurm_strerror(rc2)); rc = SLURM_ERROR; } } xfree(ops); xfree(g_context); g_context_num = -1; slurm_mutex_unlock(&g_context_lock); return rc; } extern acct_gather_energy_t *acct_gather_energy_alloc(uint16_t cnt) { return xcalloc(cnt, sizeof(struct acct_gather_energy)); } extern void acct_gather_energy_destroy(acct_gather_energy_t *energy) { xfree(energy); } extern void acct_gather_energy_pack(acct_gather_energy_t *energy, Buf buffer, uint16_t protocol_version) { if (protocol_version >= SLURM_MIN_PROTOCOL_VERSION) { if (!energy) { pack64(0, buffer); pack32(0, buffer); pack64(0, buffer); pack32(0, buffer); pack64(0, buffer); pack_time(0, buffer); return; } pack64(energy->base_consumed_energy, buffer); pack32(energy->ave_watts, buffer); pack64(energy->consumed_energy, buffer); pack32(energy->current_watts, buffer); pack64(energy->previous_consumed_energy, buffer); pack_time(energy->poll_time, buffer); } } extern int acct_gather_energy_unpack(acct_gather_energy_t **energy, Buf buffer, uint16_t protocol_version, bool need_alloc) { acct_gather_energy_t *energy_ptr; if (need_alloc) { energy_ptr = acct_gather_energy_alloc(1); *energy = energy_ptr; } else { energy_ptr = *energy; } if (protocol_version >= SLURM_MIN_PROTOCOL_VERSION) { safe_unpack64(&energy_ptr->base_consumed_energy, buffer); safe_unpack32(&energy_ptr->ave_watts, buffer); safe_unpack64(&energy_ptr->consumed_energy, buffer); safe_unpack32(&energy_ptr->current_watts, buffer); safe_unpack64(&energy_ptr->previous_consumed_energy, buffer); safe_unpack_time(&energy_ptr->poll_time, buffer); } return SLURM_SUCCESS; unpack_error: if (need_alloc) { acct_gather_energy_destroy(energy_ptr); *energy = NULL; } else memset(energy_ptr, 0, sizeof(acct_gather_energy_t)); return SLURM_ERROR; } extern int acct_gather_energy_g_update_node_energy(void) { int retval = SLURM_ERROR; if (slurm_acct_gather_energy_init() < 0) return retval; slurm_mutex_lock(&g_context_lock); for (int i = 0; i < g_context_num; i++) { if (!g_context[i]) continue; retval = (*(ops[i].update_node_energy))(); } slurm_mutex_unlock(&g_context_lock); return retval; } extern int acct_gather_energy_g_get_sum(enum acct_energy_type data_type, acct_gather_energy_t *energy) { int retval = SLURM_ERROR; static acct_gather_energy_t *e, *energy_array; if (slurm_acct_gather_energy_init() < 0) return retval; slurm_mutex_lock(&g_context_lock); if (g_context_num == 1) { retval = (*(ops[0].get_data))(data_type, energy); slurm_mutex_unlock(&g_context_lock); return retval; } energy_array = acct_gather_energy_alloc(g_context_num); for (int i = 0; i < g_context_num; i++) { if (!g_context[i]) continue; e = &energy_array[i]; retval = (*(ops[i].get_data))(data_type, e); if (retval != SLURM_SUCCESS || (e->consumed_energy == NO_VAL64)) continue; energy->base_consumed_energy += e->base_consumed_energy; energy->ave_watts += e->ave_watts; energy->consumed_energy += e->consumed_energy; energy->current_watts += e->current_watts; energy->previous_consumed_energy += e->previous_consumed_energy; /* * node poll_time is computed as the oldest poll_time of * the sensors */ if (!energy->poll_time || (energy->poll_time > e->poll_time)) energy->poll_time = e->poll_time; } slurm_mutex_unlock(&g_context_lock); acct_gather_energy_destroy(energy_array); return retval; } extern int acct_gather_energy_g_get_data(int context_id, enum acct_energy_type data_type, void *data) { int retval = SLURM_ERROR; if (slurm_acct_gather_energy_init() < 0) return retval; slurm_mutex_lock(&g_context_lock); xassert((context_id < g_context_num) && (context_id >= 0)); xassert(g_context[context_id]); retval = (*(ops[context_id].get_data))(data_type, data); slurm_mutex_unlock(&g_context_lock); return retval; } extern int acct_gather_energy_g_set_data(enum acct_energy_type data_type, void *data) { int retval = SLURM_ERROR; if (slurm_acct_gather_energy_init() < 0) return retval; slurm_mutex_lock(&g_context_lock); for (int i = 0; i < g_context_num; i++) { if (!g_context[i]) continue; retval = (*(ops[i].set_data))(data_type, data); } slurm_mutex_unlock(&g_context_lock); return retval; } extern int acct_gather_energy_startpoll(uint32_t frequency) { int retval = SLURM_SUCCESS; if (slurm_acct_gather_energy_init() < 0) return SLURM_ERROR; if (!acct_shutdown) { error("%s: poll already started!", __func__); return retval; } acct_shutdown = false; freq = frequency; if (frequency == 0) { /* don't want dynamic monitoring? */ debug2("%s: dynamic logging disabled", __func__); return retval; } /* create polling thread */ slurm_thread_create(&watch_node_thread_id, _watch_node, NULL); debug3("%s: dynamic logging enabled", __func__); return retval; } extern int acct_gather_energy_g_conf_options(s_p_options_t **full_options, int *full_options_cnt) { if (slurm_acct_gather_energy_init() < 0) return SLURM_ERROR; slurm_mutex_lock(&g_context_lock); for (int i = 0; i < g_context_num; i++) { if (!g_context[i]) continue; (*(ops[i].conf_options))(full_options, full_options_cnt); } slurm_mutex_unlock(&g_context_lock); return SLURM_SUCCESS; } extern int acct_gather_energy_g_conf_set(s_p_hashtbl_t *tbl) { if (slurm_acct_gather_energy_init() < 0) return SLURM_ERROR; slurm_mutex_lock(&g_context_lock); for (int i = 0; i < g_context_num; i++) { if (!g_context[i]) continue; (*(ops[i].conf_set))(i, tbl); } slurm_mutex_unlock(&g_context_lock); return SLURM_SUCCESS; } extern int acct_gather_energy_g_conf_values(void *data) { if (slurm_acct_gather_energy_init() < 0) return SLURM_ERROR; slurm_mutex_lock(&g_context_lock); for (int i = 0; i < g_context_num; i++) { if (!g_context[i]) continue; (*(ops[i].conf_values))(data); } slurm_mutex_unlock(&g_context_lock); return SLURM_SUCCESS; }