ports/openmpi
ports/openmpi
1
1
Fork 0
Code Releases Activity
390d29733f
openmpi/orte/runtime/orte_quit.c
Ralph Castain a200e4f865 As per the RFC, bring in the ORTE async progress code and the rewrite of OOB: 
*** THIS RFC INCLUDES A MINOR CHANGE TO THE MPI-RTE INTERFACE ***

Note: during the course of this work, it was necessary to completely separate the MPI and RTE progress engines. There were multiple places in the MPI layer where ORTE_WAIT_FOR_COMPLETION was being used. A new OMPI_WAIT_FOR_COMPLETION macro was created (defined in ompi/mca/rte/rte.h) that simply cycles across opal_progress until the provided flag becomes false. Places where the MPI layer blocked waiting for RTE to complete an event have been modified to use this macro.

***************************************************************************************

I am reissuing this RFC because of the time that has passed since its original release. Since its initial release and review, I have debugged it further to ensure it fully supports tests like loop_spawn. It therefore seems ready for merge back to the trunk. Given its prior review, I have set the timeout for one week.

The code is in  https://bitbucket.org/rhc/ompi-oob2


WHAT:    Rewrite of ORTE OOB

WHY:       Support asynchronous progress and a host of other features

WHEN:    Wed, August 21

SYNOPSIS:
The current OOB has served us well, but a number of limitations have been identified over the years. Specifically:

* it is only progressed when called via opal_progress, which can lead to hangs or recursive calls into libevent (which is not supported by that code)

* we've had issues when multiple NICs are available as the code doesn't "shift" messages between transports - thus, all nodes had to be available via the same TCP interface.

* the OOB "unloads" incoming opal_buffer_t objects during the transmission, thus preventing use of OBJ_RETAIN in the code when repeatedly sending the same message to multiple recipients

* there is no failover mechanism across NICs - if the selected NIC (or its attached switch) fails, we are forced to abort

* only one transport (i.e., component) can be "active"


The revised OOB resolves these problems:

* async progress is used for all application processes, with the progress thread blocking in the event library

* each available TCP NIC is supported by its own TCP module. The ability to asynchronously progress each module independently is provided, but not enabled by default (a runtime MCA parameter turns it "on")

* multi-address TCP NICs (e.g., a NIC with both an IPv4 and IPv6 address, or with virtual interfaces) are supported - reachability is determined by comparing the contact info for a peer against all addresses within the range covered by the address/mask pairs for the NIC.

* a message that arrives on one TCP NIC is automatically shifted to whatever NIC that is connected to the next "hop" if that peer cannot be reached by the incoming NIC. If no TCP module will reach the peer, then the OOB attempts to send the message via all other available components - if none can reach the peer, then an "error" is reported back to the RML, which then calls the errmgr for instructions.

* opal_buffer_t now conforms to standard object rules re OBJ_RETAIN as we no longer "unload" the incoming object

* NIC failure is reported to the TCP component, which then tries to resend the message across any other available TCP NIC. If that doesn't work, then the message is given back to the OOB base to try using other components. If all that fails, then the error is reported to the RML, which reports to the errmgr for instructions

* obviously from the above, multiple OOB components (e.g., TCP and UD) can be active in parallel

* the matching code has been moved to the RML (and out of the OOB/TCP component) so it is independent of transport

* routing is done by the individual OOB modules (as opposed to the RML). Thus, both routed and non-routed transports can simultaneously be active

* all blocking send/recv APIs have been removed. Everything operates asynchronously.


KNOWN LIMITATIONS:

* although provision is made for component failover as described above, the code for doing so has not been fully implemented yet. At the moment, if all connections for a given peer fail, the errmgr is notified of a "lost connection", which by default results in termination of the job if it was a lifeline

* the IPv6 code is present and compiles, but is not complete. Since the current IPv6 support in the OOB doesn't work anyway, I don't consider this a blocker

* routing is performed at the individual module level, yet the active routed component is selected on a global basis. We probably should update that to reflect that different transports may need/choose to route in different ways

* obviously, not every error path has been tested nor necessarily covered

* determining abnormal termination is more challenging than in the old code as we now potentially have multiple ways of connecting to a process. Ideally, we would declare "connection failed" when *all* transports can no longer reach the process, but that requires some additional (possibly complex) code. For now, the code replicates the old behavior only somewhat modified - i.e., if a module sees its connection fail, it checks to see if it is a lifeline. If so, it notifies the errmgr that the lifeline is lost - otherwise, it notifies the errmgr that a non-lifeline connection was lost.

* reachability is determined solely on the basis of a shared subnet address/mask - more sophisticated algorithms (e.g., the one used in the tcp btl) are required to handle routing via gateways

* the RML needs to assign sequence numbers to each message on a per-peer basis. The receiving RML will then deliver messages in order, thus preventing out-of-order messaging in the case where messages travel across different transports or a message needs to be redirected/resent due to failure of a NIC

This commit was SVN r29058.
2013-08-22 16:37:40 +00:00

391 lines
17 KiB
C
Raw Blame History

/* -*- C -*-
*
* Copyright (c) 2004-2010 The Trustees of Indiana University and Indiana
* University Research and Technology
* Corporation. All rights reserved.
* Copyright (c) 2004-2008 The University of Tennessee and The University
* of Tennessee Research Foundation. All rights
* reserved.
* Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
* University of Stuttgart. All rights reserved.
* Copyright (c) 2004-2005 The Regents of the University of California.
* All rights reserved.
* Copyright (c) 2006-2013 Cisco Systems, Inc. All rights reserved.
* Copyright (c) 2007-2009 Sun Microsystems, Inc. All rights reserved.
* Copyright (c) 2007-2013 Los Alamos National Security, LLC. All rights
* reserved.
* Copyright (c) 2012 Oak Ridge National Labs. All rights reserved.
* $COPYRIGHT$
*
* Additional copyrights may follow
*
* $HEADER$
*/
#include "orte_config.h"
#include "orte/constants.h"
#ifdef HAVE_STRING_H
#include <string.h>
#endif
#include <stdio.h>
#ifdef HAVE_UNISTD_H
#include <unistd.h>
#endif
#ifdef HAVE_SYS_PARAM_H
#include <sys/param.h>
#endif
#include <errno.h>
#include <signal.h>
#include <ctype.h>
#ifdef HAVE_SYS_TYPES_H
#include <sys/types.h>
#endif /* HAVE_SYS_TYPES_H */
#ifdef HAVE_SYS_WAIT_H
#include <sys/wait.h>
#endif /* HAVE_SYS_WAIT_H */
#ifdef HAVE_SYS_TIME_H
#include <sys/time.h>
#endif /* HAVE_SYS_TIME_H */
#include "orte/mca/plm/plm.h"
#include "orte/mca/errmgr/errmgr.h"
#include "orte/mca/routed/routed.h"
#include "orte/mca/state/state.h"
#include "orte/util/session_dir.h"
#include "orte/util/show_help.h"
#include "orte/runtime/runtime.h"
#include "orte/runtime/orte_globals.h"
#include "orte/runtime/orte_quit.h"
#include "orte/runtime/orte_locks.h"
#include "orte/runtime/orte_data_server.h"
/*
* Globals
*/
static int num_aborted = 0;
static int num_killed = 0;
static int num_failed_start = 0;
static bool errors_reported = false;
static void dump_aborted_procs(void);
void orte_quit(int fd, short args, void *cbdata)
{
orte_state_caddy_t *caddy = (orte_state_caddy_t*)cbdata;
/* cleanup */
if (NULL != caddy) {
OBJ_RELEASE(caddy);
}
/* check one-time lock to protect against "bounce" */
if (opal_atomic_trylock(&orte_quit_lock)) { /* returns 1 if already locked */
return;
}
/* if we are the hnp and haven't already reported it, then
* report any errors
*/
if (ORTE_PROC_IS_HNP && !errors_reported) {
if (0 != orte_exit_status && !orte_execute_quiet) {
errors_reported = true;
/* abnormal termination of some kind */
dump_aborted_procs();
/* If we showed more abort messages than were allowed,
show a followup message here */
if (num_failed_start > 1) {
if (orte_xml_output) {
fprintf(orte_xml_fp, "<stderr>");
}
fprintf(orte_xml_fp, "%d total process%s failed to start",
num_failed_start, ((num_failed_start > 1) ? "es" : ""));
if (orte_xml_output) {
fprintf(orte_xml_fp, "&#010;</stderr>");
}
fprintf(orte_xml_fp, "\n");
}
if (num_aborted > 1) {
if (orte_xml_output) {
fprintf(orte_xml_fp, "<stderr>");
}
fprintf(orte_xml_fp, "%d total process%s aborted",
num_aborted, ((num_aborted > 1) ? "es" : ""));
if (orte_xml_output) {
fprintf(orte_xml_fp, "&#010;</stderr>");
}
fprintf(orte_xml_fp, "\n");
}
if (num_killed > 1) {
if (orte_xml_output) {
fprintf(orte_xml_fp, "<stderr>");
}
fprintf(orte_xml_fp, "%d total process%s killed (some possibly by %s during cleanup)",
num_killed, ((num_killed > 1) ? "es" : ""), orte_basename);
if (orte_xml_output) {
fprintf(orte_xml_fp, "&#010;</stderr>");
}
fprintf(orte_xml_fp, "\n");
}
}
}
/* flag that the event lib should no longer be looped
* so we will exit
*/
orte_event_base_active = false;
/* break out of the event loop */
opal_event_base_loopbreak(orte_event_base);
}
/*
* On abnormal termination - dump the
* exit status of the aborted procs.
*/
static void dump_aborted_procs(void)
{
orte_std_cntr_t i, n;
orte_proc_t *proc, *pptr;
orte_app_context_t *approc;
orte_job_t *job;
orte_node_t *node;
/* find the job that caused the problem - be sure to start the loop
* at 1 as the daemons are in 0 and will clearly be "running", so no
* point in checking them
*/
for (n=1; n < orte_job_data->size; n++) {
if (NULL == (job = (orte_job_t*)opal_pointer_array_get_item(orte_job_data, n))) {
/* the array is no longer left-justified, so we have to continue */
continue;
}
if (ORTE_JOB_STATE_UNDEF != job->state &&
ORTE_JOB_STATE_INIT != job->state &&
ORTE_JOB_STATE_RUNNING != job->state &&
ORTE_JOB_STATE_TERMINATED != job->state &&
ORTE_JOB_STATE_ABORT_ORDERED != job->state) {
/* this is a guilty party */
proc = job->aborted_proc;
/* cycle through and count the number that were killed or aborted */
for (i=0; i < job->procs->size; i++) {
if (NULL == (pptr = (orte_proc_t*)opal_pointer_array_get_item(job->procs, i))) {
/* array is left-justfied - we are done */
continue;
}
if (ORTE_PROC_STATE_FAILED_TO_START == pptr->state ||
ORTE_PROC_STATE_FAILED_TO_LAUNCH == pptr->state) {
++num_failed_start;
} else if (ORTE_PROC_STATE_ABORTED == pptr->state) {
++num_aborted;
} else if (ORTE_PROC_STATE_ABORTED_BY_SIG == pptr->state) {
++num_killed;
} else if (ORTE_PROC_STATE_SENSOR_BOUND_EXCEEDED == pptr->state) {
++num_killed;
}
}
if (NULL == proc) {
continue;
}
approc = (orte_app_context_t*)opal_pointer_array_get_item(job->apps, proc->app_idx);
node = proc->node;
if (ORTE_JOB_STATE_FAILED_TO_START == job->state ||
ORTE_JOB_STATE_FAILED_TO_LAUNCH == job->state) {
if (NULL == proc) {
orte_show_help("help-orterun.txt", "orterun:proc-failed-to-start-no-status-no-node", true,
orte_basename);
return;
}
switch (proc->exit_code) {
case ORTE_ERR_SILENT:
/* say nothing - it was already reported */
break;
case ORTE_ERR_SYS_LIMITS_PIPES:
orte_show_help("help-orterun.txt", "orterun:sys-limit-pipe", true,
orte_basename, proc->node->name,
(unsigned long)proc->name.vpid);
break;
case ORTE_ERR_PIPE_SETUP_FAILURE:
orte_show_help("help-orterun.txt", "orterun:pipe-setup-failure", true,
orte_basename, proc->node->name,
(unsigned long)proc->name.vpid);
break;
case ORTE_ERR_SYS_LIMITS_CHILDREN:
orte_show_help("help-orterun.txt", "orterun:sys-limit-children", true,
orte_basename, proc->node->name,
(unsigned long)proc->name.vpid);
break;
case ORTE_ERR_FAILED_GET_TERM_ATTRS:
orte_show_help("help-orterun.txt", "orterun:failed-term-attrs", true,
orte_basename, proc->node->name,
(unsigned long)proc->name.vpid);
break;
case ORTE_ERR_WDIR_NOT_FOUND:
orte_show_help("help-orterun.txt", "orterun:wdir-not-found", true,
orte_basename, approc->cwd,
proc->node->name, (unsigned long)proc->name.vpid);
break;
case ORTE_ERR_EXE_NOT_FOUND:
orte_show_help("help-orterun.txt", "orterun:exe-not-found", true,
orte_basename,
(unsigned long)proc->name.vpid,
orte_basename,
orte_basename,
proc->node->name,
approc->app);
break;
case ORTE_ERR_EXE_NOT_ACCESSIBLE:
orte_show_help("help-orterun.txt", "orterun:exe-not-accessible", true,
orte_basename, approc->app, proc->node->name,
(unsigned long)proc->name.vpid);
break;
case ORTE_ERR_MULTIPLE_AFFINITIES:
orte_show_help("help-orterun.txt",
"orterun:multiple-paffinity-schemes", true, NULL);
break;
case ORTE_ERR_TOPO_SLOT_LIST_NOT_SUPPORTED:
orte_show_help("help-orterun.txt",
"orterun:topo-not-supported",
true, orte_process_info.nodename, "rankfile containing a slot_list of ",
NULL, approc->app);
break;
case ORTE_ERR_INVALID_NODE_RANK:
orte_show_help("help-orterun.txt",
"orterun:invalid-node-rank", true);
break;
case ORTE_ERR_INVALID_LOCAL_RANK:
orte_show_help("help-orterun.txt",
"orterun:invalid-local-rank", true);
break;
case ORTE_ERR_NOT_ENOUGH_CORES:
orte_show_help("help-orterun.txt",
"orterun:not-enough-resources", true,
"sockets", node->name,
"bind-to-core", approc->app);
break;
case ORTE_ERR_TOPO_CORE_NOT_SUPPORTED:
orte_show_help("help-orterun.txt",
"orterun:topo-not-supported",
true, node->name, "bind-to-core", "",
approc->app);
break;
case ORTE_ERR_INVALID_PHYS_CPU:
orte_show_help("help-orterun.txt",
"orterun:invalid-phys-cpu", true);
break;
case ORTE_ERR_NOT_ENOUGH_SOCKETS:
orte_show_help("help-orterun.txt",
"orterun:not-enough-resources", true,
"sockets", node->name,
"bind-to-socket", approc->app);
break;
case ORTE_ERR_TOPO_SOCKET_NOT_SUPPORTED:
orte_show_help("help-orterun.txt",
"orterun:topo-not-supported",
true, node->name, "bind-to-socket", "",
approc->app);
break;
case ORTE_ERR_MODULE_NOT_FOUND:
orte_show_help("help-orterun.txt",
"orterun:paffinity-missing-module",
true, node->name);
break;
case ORTE_ERR_SLOT_LIST_RANGE:
orte_show_help("help-orterun.txt",
"orterun:invalid-slot-list-range",
true, node->name, NULL);
break;
case ORTE_ERR_PIPE_READ_FAILURE:
orte_show_help("help-orterun.txt", "orterun:pipe-read-failure", true,
orte_basename, node->name, (unsigned long)proc->name.vpid);
break;
case ORTE_ERR_SOCKET_NOT_AVAILABLE:
orte_show_help("help-orterun.txt", "orterun:proc-socket-not-avail", true,
orte_basename, ORTE_ERROR_NAME(proc->exit_code), node->name,
(unsigned long)proc->name.vpid);
break;
default:
if (0 != proc->exit_code) {
orte_show_help("help-orterun.txt", "orterun:proc-failed-to-start", true,
orte_basename, ORTE_ERROR_NAME(proc->exit_code), node->name,
(unsigned long)proc->name.vpid);
} else {
orte_show_help("help-orterun.txt", "orterun:proc-failed-to-start-no-status", true,
orte_basename, node->name);
}
break;
}
} else if (ORTE_JOB_STATE_ABORTED == job->state) {
if (NULL == proc) {
orte_show_help("help-orterun.txt", "orterun:proc-aborted-unknown", true,
orte_basename);
} else {
orte_show_help("help-orterun.txt", "orterun:proc-ordered-abort", true,
orte_basename, (unsigned long)proc->name.vpid, (unsigned long)proc->pid,
node->name, orte_basename);
}
} else if (ORTE_JOB_STATE_ABORTED_BY_SIG == job->state) { /* aborted by signal */
if (NULL == proc) {
orte_show_help("help-orterun.txt", "orterun:proc-aborted-signal-unknown", true,
orte_basename);
} else {
#ifdef HAVE_STRSIGNAL
if (NULL != strsignal(WTERMSIG(proc->exit_code))) {
orte_show_help("help-orterun.txt", "orterun:proc-aborted-strsignal", true,
orte_basename, (unsigned long)proc->name.vpid, (unsigned long)proc->pid,
node->name, WTERMSIG(proc->exit_code),
strsignal(WTERMSIG(proc->exit_code)));
} else {
#endif
orte_show_help("help-orterun.txt", "orterun:proc-aborted", true,
orte_basename, (unsigned long)proc->name.vpid, (unsigned long)proc->pid,
node->name, WTERMSIG(proc->exit_code));
#ifdef HAVE_STRSIGNAL
}
#endif
}
} else if (ORTE_JOB_STATE_ABORTED_WO_SYNC == job->state) { /* proc exited w/o finalize */
if (NULL == proc) {
orte_show_help("help-orterun.txt", "orterun:proc-exit-no-sync-unknown", true,
orte_basename, orte_basename);
} else {
orte_show_help("help-orterun.txt", "orterun:proc-exit-no-sync", true,
orte_basename, (unsigned long)proc->name.vpid, (unsigned long)proc->pid,
node->name, orte_basename, orte_basename);
}
} else if (ORTE_JOB_STATE_COMM_FAILED == job->state) {
orte_show_help("help-orterun.txt", "orterun:proc-comm-failed", true,
ORTE_NAME_PRINT(ORTE_PROC_MY_NAME),
ORTE_NAME_PRINT(&proc->name), node->name);
} else if (ORTE_JOB_STATE_SENSOR_BOUND_EXCEEDED == job->state) {
switch (proc->exit_code) {
case ORTE_ERR_MEM_LIMIT_EXCEEDED:
orte_show_help("help-orterun.txt", "orterun:proc-mem-exceeded", true,
ORTE_NAME_PRINT(&proc->name), node->name);
break;
case ORTE_ERR_PROC_STALLED:
orte_show_help("help-orterun.txt", "orterun:proc-stalled", true);
break;
default:
orte_show_help("help-orterun.txt", "orterun:proc-sensor-exceeded", true);
break;
}
} else if (ORTE_JOB_STATE_HEARTBEAT_FAILED == job->state) {
orte_show_help("help-orterun.txt", "orterun:proc-heartbeat-failed", true,
orte_basename, ORTE_NAME_PRINT(&proc->name), node->name);
} else if (orte_abort_non_zero_exit &&
ORTE_JOB_STATE_NON_ZERO_TERM == job->state) {
orte_show_help("help-orterun.txt", "orterun:non-zero-exit", true,
orte_basename, ORTE_NAME_PRINT(&proc->name), proc->exit_code);
}
return;
}
}
}
View Git Blame Copy Permalink