/* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*- * vim:expandtab:shiftwidth=8:tabstop=8: * * Portal-RPC reconnection and replay operations, for use in recovery. * * Copyright (c) 2003 Cluster File Systems, Inc. * Authors: Phil Schwan * Mike Shaver * * This file is part of the Lustre file system, http://www.lustre.org * Lustre is a trademark of Cluster File Systems, Inc. * * You may have signed or agreed to another license before downloading * this software. If so, you are bound by the terms and conditions * of that agreement, and the following does not apply to you. See the * LICENSE file included with this distribution for more information. * * If you did not agree to a different license, then this copy of Lustre * is open source software; you can redistribute it and/or modify it * under the terms of version 2 of the GNU General Public License as * published by the Free Software Foundation. * * In either case, Lustre 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 * license text for more details. */ #ifndef __KERNEL__ #include #else #define DEBUG_SUBSYSTEM S_RPC #endif #include #include #include "ptlrpc_internal.h" struct semaphore pinger_sem; static CFS_LIST_HEAD(pinger_imports); int ptlrpc_ping(struct obd_import *imp) { struct ptlrpc_request *req; ENTRY; req = ptlrpc_request_alloc_pack(imp, &RQF_OBD_PING, LUSTRE_OBD_VERSION, OBD_PING); if (req == NULL) { CERROR("OOM trying to ping %s->%s\n", imp->imp_obd->obd_uuid.uuid, obd2cli_tgt(imp->imp_obd)); RETURN(-ENOMEM); } DEBUG_REQ(D_INFO, req, "pinging %s->%s", imp->imp_obd->obd_uuid.uuid, obd2cli_tgt(imp->imp_obd)); req->rq_no_resend = req->rq_no_delay = 1; ptlrpc_request_set_replen(req); req->rq_timeout = PING_INTERVAL; ptlrpcd_add_req(req); RETURN(0); } void ptlrpc_update_next_ping(struct obd_import *imp) { #ifdef ENABLE_PINGER imp->imp_next_ping = cfs_time_shift( (imp->imp_state == LUSTRE_IMP_DISCON ? RECONNECT_INTERVAL : PING_INTERVAL)); #endif /* ENABLE_PINGER */ } void ptlrpc_ping_import_soon(struct obd_import *imp) { imp->imp_next_ping = cfs_time_current(); } static inline int ptlrpc_next_reconnect(struct obd_import *imp) { if (imp->imp_server_timeout) return cfs_time_shift(obd_timeout / 2); else return cfs_time_shift(obd_timeout); } static atomic_t suspend_timeouts = ATOMIC_INIT(0); static cfs_time_t suspend_wakeup_time = 0; #ifdef __KERNEL__ static wait_queue_head_t suspend_timeouts_waitq; #endif cfs_time_t ptlrpc_suspend_wakeup_time(void) { return suspend_wakeup_time; } void ptlrpc_deactivate_timeouts(struct obd_import *imp) { /*XXX: disabled for now, will be replaced by adaptive timeouts */ #if 0 if (imp->imp_no_timeout) return; imp->imp_no_timeout = 1; atomic_inc(&suspend_timeouts); CDEBUG(D_HA|D_WARNING, "deactivate timeouts %u\n", atomic_read(&suspend_timeouts)); #endif } void ptlrpc_activate_timeouts(struct obd_import *imp) { /*XXX: disabled for now, will be replaced by adaptive timeouts */ #if 0 if (!imp->imp_no_timeout) return; imp->imp_no_timeout = 0; LASSERT(atomic_read(&suspend_timeouts) > 0); if (atomic_dec_and_test(&suspend_timeouts)) { suspend_wakeup_time = cfs_time_current(); wake_up(&suspend_timeouts_waitq); } CDEBUG(D_HA|D_WARNING, "activate timeouts %u\n", atomic_read(&suspend_timeouts)); #endif } int ptlrpc_check_suspend(void) { if (atomic_read(&suspend_timeouts)) return 1; return 0; } int ptlrpc_check_and_wait_suspend(struct ptlrpc_request *req) { struct l_wait_info lwi; if (atomic_read(&suspend_timeouts)) { DEBUG_REQ(D_NET, req, "-- suspend %d regular timeout", atomic_read(&suspend_timeouts)); lwi = LWI_INTR(NULL, NULL); l_wait_event(suspend_timeouts_waitq, atomic_read(&suspend_timeouts) == 0, &lwi); DEBUG_REQ(D_NET, req, "-- recharge regular timeout"); return 1; } return 0; } #ifdef __KERNEL__ static void ptlrpc_pinger_process_import(struct obd_import *imp, unsigned long this_ping) { int force, level; spin_lock(&imp->imp_lock); level = imp->imp_state; force = imp->imp_force_verify; if (force) imp->imp_force_verify = 0; spin_unlock(&imp->imp_lock); CDEBUG(level == LUSTRE_IMP_FULL ? D_INFO : D_HA, "level %s/%u force %u deactive %u pingable %u\n", ptlrpc_import_state_name(level), level, force, imp->imp_deactive, imp->imp_pingable); if (cfs_time_aftereq(imp->imp_next_ping - 5 * CFS_TICK, this_ping) && force == 0) return; if (level == LUSTRE_IMP_DISCON && !imp->imp_deactive) { /* wait at least a timeout before trying recovery again */ imp->imp_next_ping = ptlrpc_next_reconnect(imp); ptlrpc_initiate_recovery(imp); } else if (level != LUSTRE_IMP_FULL || imp->imp_obd->obd_no_recov || imp->imp_deactive) { CDEBUG(D_HA, "not pinging %s (in recovery " " or recovery disabled: %s)\n", obd2cli_tgt(imp->imp_obd), ptlrpc_import_state_name(level)); } else if (imp->imp_pingable || force) { ptlrpc_ping(imp); } } static int ptlrpc_pinger_main(void *arg) { struct ptlrpc_svc_data *data = (struct ptlrpc_svc_data *)arg; struct ptlrpc_thread *thread = data->thread; ENTRY; cfs_daemonize(data->name); /* Record that the thread is running */ thread->t_flags = SVC_RUNNING; cfs_waitq_signal(&thread->t_ctl_waitq); /* And now, loop forever, pinging as needed. */ while (1) { cfs_time_t this_ping = cfs_time_current(); struct l_wait_info lwi; cfs_duration_t time_to_next_ping; struct list_head *iter; mutex_down(&pinger_sem); list_for_each(iter, &pinger_imports) { struct obd_import *imp = list_entry(iter, struct obd_import, imp_pinger_chain); ptlrpc_pinger_process_import(imp, this_ping); /* obd_timeout might have changed */ if (imp->imp_pingable && imp->imp_next_ping && cfs_time_after(imp->imp_next_ping, cfs_time_add(this_ping, cfs_time_seconds(PING_INTERVAL)))) ptlrpc_update_next_ping(imp); } mutex_up(&pinger_sem); /* update memory usage info */ obd_update_maxusage(); /* Wait until the next ping time, or until we're stopped. */ time_to_next_ping = cfs_time_sub(cfs_time_add(this_ping, cfs_time_seconds(PING_INTERVAL)), cfs_time_current()); /* The ping sent by ptlrpc_send_rpc may get sent out say .01 second after this. ptlrpc_pinger_eending_on_import will then set the next ping time to next_ping + .01 sec, which means we will SKIP the next ping at next_ping, and the ping will get sent 2 timeouts from now! Beware. */ CDEBUG(D_INFO, "next ping in "CFS_DURATION_T" ("CFS_TIME_T")\n", time_to_next_ping, cfs_time_add(this_ping, cfs_time_seconds(PING_INTERVAL))); if (time_to_next_ping > 0) { lwi = LWI_TIMEOUT(max_t(cfs_duration_t, time_to_next_ping, cfs_time_seconds(1)), NULL, NULL); l_wait_event(thread->t_ctl_waitq, thread->t_flags & (SVC_STOPPING|SVC_EVENT), &lwi); if (thread->t_flags & SVC_STOPPING) { thread->t_flags &= ~SVC_STOPPING; EXIT; break; } else if (thread->t_flags & SVC_EVENT) { /* woken after adding import to reset timer */ thread->t_flags &= ~SVC_EVENT; } } } thread->t_flags = SVC_STOPPED; cfs_waitq_signal(&thread->t_ctl_waitq); CDEBUG(D_NET, "pinger thread exiting, process %d\n", cfs_curproc_pid()); return 0; } static struct ptlrpc_thread *pinger_thread = NULL; int ptlrpc_start_pinger(void) { struct l_wait_info lwi = { 0 }; struct ptlrpc_svc_data d; int rc; #ifndef ENABLE_PINGER return 0; #endif ENTRY; if (pinger_thread != NULL) RETURN(-EALREADY); OBD_ALLOC_PTR(pinger_thread); if (pinger_thread == NULL) RETURN(-ENOMEM); cfs_waitq_init(&pinger_thread->t_ctl_waitq); cfs_waitq_init(&suspend_timeouts_waitq); d.name = "ll_ping"; d.thread = pinger_thread; /* CLONE_VM and CLONE_FILES just avoid a needless copy, because we * just drop the VM and FILES in ptlrpc_daemonize() right away. */ rc = cfs_kernel_thread(ptlrpc_pinger_main, &d, CLONE_VM | CLONE_FILES); if (rc < 0) { CERROR("cannot start thread: %d\n", rc); OBD_FREE(pinger_thread, sizeof(*pinger_thread)); pinger_thread = NULL; RETURN(rc); } l_wait_event(pinger_thread->t_ctl_waitq, pinger_thread->t_flags & SVC_RUNNING, &lwi); RETURN(0); } int ptlrpc_stop_pinger(void) { struct l_wait_info lwi = { 0 }; int rc = 0; #ifndef ENABLE_PINGER return 0; #endif ENTRY; if (pinger_thread == NULL) RETURN(-EALREADY); mutex_down(&pinger_sem); pinger_thread->t_flags = SVC_STOPPING; cfs_waitq_signal(&pinger_thread->t_ctl_waitq); mutex_up(&pinger_sem); l_wait_event(pinger_thread->t_ctl_waitq, (pinger_thread->t_flags & SVC_STOPPED), &lwi); OBD_FREE_PTR(pinger_thread); pinger_thread = NULL; RETURN(rc); } void ptlrpc_pinger_sending_on_import(struct obd_import *imp) { ptlrpc_update_next_ping(imp); } int ptlrpc_pinger_add_import(struct obd_import *imp) { ENTRY; if (!list_empty(&imp->imp_pinger_chain)) RETURN(-EALREADY); mutex_down(&pinger_sem); CDEBUG(D_HA, "adding pingable import %s->%s\n", imp->imp_obd->obd_uuid.uuid, obd2cli_tgt(imp->imp_obd)); /* if we add to pinger we want recovery on this import */ imp->imp_obd->obd_no_recov = 0; ptlrpc_update_next_ping(imp); /* XXX sort, blah blah */ list_add_tail(&imp->imp_pinger_chain, &pinger_imports); class_import_get(imp); ptlrpc_pinger_wake_up(); mutex_up(&pinger_sem); RETURN(0); } int ptlrpc_pinger_del_import(struct obd_import *imp) { ENTRY; if (list_empty(&imp->imp_pinger_chain)) RETURN(-ENOENT); mutex_down(&pinger_sem); list_del_init(&imp->imp_pinger_chain); CDEBUG(D_HA, "removing pingable import %s->%s\n", imp->imp_obd->obd_uuid.uuid, obd2cli_tgt(imp->imp_obd)); /* if we remove from pinger we don't want recovery on this import */ imp->imp_obd->obd_no_recov = 1; class_import_put(imp); mutex_up(&pinger_sem); RETURN(0); } void ptlrpc_pinger_wake_up() { #ifdef ENABLE_PINGER pinger_thread->t_flags |= SVC_EVENT; cfs_waitq_signal(&pinger_thread->t_ctl_waitq); #endif } /* Ping evictor thread */ #define PET_READY 1 #define PET_TERMINATE 2 static int pet_refcount = 0; static int pet_state; static wait_queue_head_t pet_waitq; static struct obd_export *pet_exp = NULL; static spinlock_t pet_lock = SPIN_LOCK_UNLOCKED; int ping_evictor_wake(struct obd_export *exp) { spin_lock(&pet_lock); if (pet_exp || (pet_state != PET_READY)) { /* eventually the new obd will call here again. */ spin_unlock(&pet_lock); return 1; } /* We have to make sure the obd isn't destroyed between now and when * the ping evictor runs. We'll take a reference here, and drop it * when we finish in the evictor. We don't really care about this * export in particular; we just need one to keep the obd alive. */ pet_exp = class_export_get(exp); spin_unlock(&pet_lock); wake_up(&pet_waitq); return 0; } static int ping_evictor_main(void *arg) { struct obd_device *obd; struct obd_export *exp; struct l_wait_info lwi = { 0 }; time_t expire_time; ENTRY; ptlrpc_daemonize("ll_evictor"); CDEBUG(D_HA, "Starting Ping Evictor\n"); pet_exp = NULL; pet_state = PET_READY; while (1) { l_wait_event(pet_waitq, pet_exp || (pet_state == PET_TERMINATE), &lwi); if (pet_state == PET_TERMINATE) break; /* we only get here if pet_exp != NULL, and the end of this * loop is the only place which sets it NULL again, so lock * is not strictly necessary. */ spin_lock(&pet_lock); obd = pet_exp->exp_obd; spin_unlock(&pet_lock); expire_time = cfs_time_current_sec() - (3 * obd_timeout / 2); CDEBUG(D_HA, "evicting all exports of obd %s older than %ld\n", obd->obd_name, expire_time); /* Exports can't be deleted out of the list while we hold * the obd lock (class_unlink_export), which means we can't * lose the last ref on the export. If they've already been * removed from the list, we won't find them here. */ spin_lock(&obd->obd_dev_lock); while (!list_empty(&obd->obd_exports_timed)) { exp = list_entry(obd->obd_exports_timed.next, struct obd_export,exp_obd_chain_timed); if (expire_time > exp->exp_last_request_time) { class_export_get(exp); spin_unlock(&obd->obd_dev_lock); LCONSOLE_WARN("%s: haven't heard from client %s" " (at %s) in %ld seconds. I think" " it's dead, and I am evicting" " it. exp %p, cur %ld expire %ld" " last %ld\n", obd->obd_name, obd_uuid2str(&exp->exp_client_uuid), obd_export_nid2str(exp), (long)(cfs_time_current_sec() - exp->exp_last_request_time), exp, (long)cfs_time_current_sec(), (long)expire_time, (long)exp->exp_last_request_time); CDEBUG(D_HA, "Last request was at %ld\n", exp->exp_last_request_time); class_fail_export(exp); class_export_put(exp); spin_lock(&obd->obd_dev_lock); } else { /* List is sorted, so everyone below is ok */ break; } } spin_unlock(&obd->obd_dev_lock); class_export_put(pet_exp); spin_lock(&pet_lock); pet_exp = NULL; spin_unlock(&pet_lock); } CDEBUG(D_HA, "Exiting Ping Evictor\n"); RETURN(0); } void ping_evictor_start(void) { int rc; if (++pet_refcount > 1) return; init_waitqueue_head(&pet_waitq); rc = cfs_kernel_thread(ping_evictor_main, NULL, CLONE_VM | CLONE_FILES); if (rc < 0) { pet_refcount--; CERROR("Cannot start ping evictor thread: %d\n", rc); } } EXPORT_SYMBOL(ping_evictor_start); void ping_evictor_stop(void) { if (--pet_refcount > 0) return; pet_state = PET_TERMINATE; wake_up(&pet_waitq); } EXPORT_SYMBOL(ping_evictor_stop); #else /* !__KERNEL__ */ /* XXX * the current implementation of pinger in liblustre is not optimized */ #ifdef ENABLE_PINGER static struct pinger_data { int pd_recursion; cfs_time_t pd_this_ping; /* jiffies */ cfs_time_t pd_next_ping; /* jiffies */ struct ptlrpc_request_set *pd_set; } pinger_args; static int pinger_check_rpcs(void *arg) { cfs_time_t curtime = cfs_time_current(); struct ptlrpc_request *req; struct ptlrpc_request_set *set; struct list_head *iter; struct pinger_data *pd = &pinger_args; int rc; /* prevent recursion */ if (pd->pd_recursion++) { CDEBUG(D_HA, "pinger: recursion! quit\n"); LASSERT(pd->pd_set); pd->pd_recursion--; return 0; } /* have we reached ping point? */ if (!pd->pd_set && time_before(curtime, pd->pd_next_ping)) { pd->pd_recursion--; return 0; } /* if we have rpc_set already, continue processing it */ if (pd->pd_set) { LASSERT(pd->pd_this_ping); set = pd->pd_set; goto do_check_set; } pd->pd_this_ping = curtime; pd->pd_set = ptlrpc_prep_set(); if (pd->pd_set == NULL) goto out; set = pd->pd_set; /* add rpcs into set */ mutex_down(&pinger_sem); list_for_each(iter, &pinger_imports) { struct obd_import *imp = list_entry(iter, struct obd_import, imp_pinger_chain); int generation, level; if (cfs_time_aftereq(pd->pd_this_ping, imp->imp_next_ping - 5 * CFS_TICK)) { /* Add a ping. */ spin_lock(&imp->imp_lock); generation = imp->imp_generation; level = imp->imp_state; spin_unlock(&imp->imp_lock); if (level != LUSTRE_IMP_FULL) { CDEBUG(D_HA, "not pinging %s (in recovery)\n", obd2cli_tgt(imp->imp_obd)); continue; } req = ptlrpc_request_alloc_pack(imp, &RQF_OBD_PING, LUSTRE_OBD_VERSION, OBD_PING); if (req == NULL) { CERROR("OOM trying to ping %s->%s\n", imp->imp_obd->obd_uuid.uuid, obd2cli_tgt(imp->imp_obd)); break; } req->rq_no_resend = 1; ptlrpc_request_set_replen(req); req->rq_send_state = LUSTRE_IMP_FULL; req->rq_phase = RQ_PHASE_RPC; req->rq_import_generation = generation; ptlrpc_set_add_req(set, req); } else { CDEBUG(D_INFO, "don't need to ping %s ("CFS_TIME_T " > "CFS_TIME_T")\n", obd2cli_tgt(imp->imp_obd), imp->imp_next_ping, pd->pd_this_ping); } } pd->pd_this_ping = curtime; mutex_up(&pinger_sem); /* Might be empty, that's OK. */ if (set->set_remaining == 0) CDEBUG(D_RPCTRACE, "nothing to ping\n"); list_for_each(iter, &set->set_requests) { struct ptlrpc_request *req = list_entry(iter, struct ptlrpc_request, rq_set_chain); DEBUG_REQ(D_RPCTRACE, req, "pinging %s->%s", req->rq_import->imp_obd->obd_uuid.uuid, obd2cli_tgt(req->rq_import->imp_obd)); (void)ptl_send_rpc(req, 0); } do_check_set: rc = ptlrpc_check_set(set); /* not finished, and we are not expired, simply return */ if (!rc && cfs_time_before(curtime, cfs_time_add(pd->pd_this_ping, cfs_time_seconds(PING_INTERVAL)))) { CDEBUG(D_RPCTRACE, "not finished, but also not expired\n"); pd->pd_recursion--; return 0; } /* Expire all the requests that didn't come back. */ mutex_down(&pinger_sem); list_for_each(iter, &set->set_requests) { req = list_entry(iter, struct ptlrpc_request, rq_set_chain); if (req->rq_phase == RQ_PHASE_COMPLETE) continue; req->rq_phase = RQ_PHASE_COMPLETE; atomic_dec(&req->rq_import->imp_inflight); set->set_remaining--; /* If it was disconnected, don't sweat it. */ if (list_empty(&req->rq_import->imp_pinger_chain)) { ptlrpc_unregister_reply(req); continue; } CDEBUG(D_RPCTRACE, "pinger initiate expire_one_request\n"); ptlrpc_expire_one_request(req); } mutex_up(&pinger_sem); ptlrpc_set_destroy(set); pd->pd_set = NULL; out: pd->pd_next_ping = cfs_time_add(pd->pd_this_ping, cfs_time_seconds(PING_INTERVAL)); pd->pd_this_ping = 0; /* XXX for debug */ CDEBUG(D_INFO, "finished a round ping\n"); pd->pd_recursion--; return 0; } static void *pinger_callback = NULL; #endif /* ENABLE_PINGER */ int ptlrpc_start_pinger(void) { #ifdef ENABLE_PINGER memset(&pinger_args, 0, sizeof(pinger_args)); pinger_callback = liblustre_register_wait_callback("pinger_check_rpcs", &pinger_check_rpcs, &pinger_args); #endif return 0; } int ptlrpc_stop_pinger(void) { #ifdef ENABLE_PINGER if (pinger_callback) liblustre_deregister_wait_callback(pinger_callback); #endif return 0; } void ptlrpc_pinger_sending_on_import(struct obd_import *imp) { #ifdef ENABLE_PINGER mutex_down(&pinger_sem); ptlrpc_update_next_ping(imp); if (pinger_args.pd_set == NULL && time_before(imp->imp_next_ping, pinger_args.pd_next_ping)) { CDEBUG(D_HA, "set next ping to "CFS_TIME_T"(cur "CFS_TIME_T")\n", imp->imp_next_ping, cfs_time_current()); pinger_args.pd_next_ping = imp->imp_next_ping; } mutex_up(&pinger_sem); #endif } int ptlrpc_pinger_add_import(struct obd_import *imp) { ENTRY; if (!list_empty(&imp->imp_pinger_chain)) RETURN(-EALREADY); CDEBUG(D_HA, "adding pingable import %s->%s\n", imp->imp_obd->obd_uuid.uuid, obd2cli_tgt(imp->imp_obd)); ptlrpc_pinger_sending_on_import(imp); mutex_down(&pinger_sem); list_add_tail(&imp->imp_pinger_chain, &pinger_imports); class_import_get(imp); mutex_up(&pinger_sem); RETURN(0); } int ptlrpc_pinger_del_import(struct obd_import *imp) { ENTRY; if (list_empty(&imp->imp_pinger_chain)) RETURN(-ENOENT); mutex_down(&pinger_sem); list_del_init(&imp->imp_pinger_chain); CDEBUG(D_HA, "removing pingable import %s->%s\n", imp->imp_obd->obd_uuid.uuid, obd2cli_tgt(imp->imp_obd)); class_import_put(imp); mutex_up(&pinger_sem); RETURN(0); } void ptlrpc_pinger_wake_up() { #ifdef ENABLE_PINGER /* XXX force pinger to run, if needed */ struct obd_import *imp; ENTRY; list_for_each_entry(imp, &pinger_imports, imp_pinger_chain) { CDEBUG(D_RPCTRACE, "checking import %s->%s\n", imp->imp_obd->obd_uuid.uuid, obd2cli_tgt(imp->imp_obd)); #ifdef ENABLE_LIBLUSTRE_RECOVERY if (imp->imp_state == LUSTRE_IMP_DISCON && !imp->imp_deactive) #else /*XXX only recover for the initial connection */ if (!lustre_handle_is_used(&imp->imp_remote_handle) && imp->imp_state == LUSTRE_IMP_DISCON && !imp->imp_deactive) #endif ptlrpc_initiate_recovery(imp); else if (imp->imp_state != LUSTRE_IMP_FULL) CDEBUG(D_HA, "Refused to recover import %s->%s " "state %d, deactive %d\n", imp->imp_obd->obd_uuid.uuid, obd2cli_tgt(imp->imp_obd), imp->imp_state, imp->imp_deactive); } EXIT; #endif } #endif /* !__KERNEL__ */