Add the process PLS. It's only intended for Windows users.
This commit was SVN r13051.
Этот коммит содержится в:
родитель
950149ec50
Коммит
c7da2b0a9a
0
orte/mca/pls/process/.ompi_ignore
Обычный файл
0
orte/mca/pls/process/.ompi_ignore
Обычный файл
54
orte/mca/pls/process/help-pls-process.txt
Обычный файл
54
orte/mca/pls/process/help-pls-process.txt
Обычный файл
@ -0,0 +1,54 @@
|
||||
# -*- text -*-
|
||||
#
|
||||
# Copyright (c) 2004-2006 The University of Tennessee and The University
|
||||
# of Tennessee Research Foundation. All rights
|
||||
# reserved.
|
||||
# $COPYRIGHT$
|
||||
#
|
||||
# Additional copyrights may follow
|
||||
#
|
||||
# $HEADER$
|
||||
#
|
||||
# This is the US/English general help file for Open RTE's orterun.
|
||||
#
|
||||
[no-local-orted]
|
||||
The process PLS component was not able to find the executable "orted" in
|
||||
your PATH or in the directory where Open MPI was initially installed,
|
||||
and therefore cannot continue.
|
||||
|
||||
For reference, your current PATH is:
|
||||
|
||||
%s
|
||||
|
||||
We also looked for orted in the following directory:
|
||||
|
||||
%s
|
||||
|
||||
[multiple-prefixes]
|
||||
Specified multiple application contexts using different
|
||||
settings for --prefix. Care should be taken, that corresponding
|
||||
processes are mapped to different nodes. Having multiple prefixes
|
||||
per node is not allowed.
|
||||
|
||||
The previously set prefix was
|
||||
%s
|
||||
|
||||
the prefix to be set overriding:
|
||||
%s
|
||||
|
||||
[concurrency-less-than-zero]
|
||||
The value of the MCA parameter "pls_process_num_concurrent" is less than
|
||||
or equal to zero (%d). This parameter is used to determine how many
|
||||
remote agents (typically process or ssh) to invoke concurrently while
|
||||
launching parallel jobs.
|
||||
|
||||
This value has automatically be reset to 1; processing will continue.
|
||||
|
||||
[deadlock-params]
|
||||
The process launcher has been given a number of %d concurrent daemons to
|
||||
launch and is in a debug-daemons option. However, the total number of
|
||||
daemons to launch (%d) is greater than this value. This is a scenario that
|
||||
will cause the system to deadlock.
|
||||
|
||||
To avoid deadlock, either increase the number of concurrent daemons, or
|
||||
remove the debug-daemons flag.
|
90
orte/mca/pls/process/pls_process.h
Обычный файл
90
orte/mca/pls/process/pls_process.h
Обычный файл
@ -0,0 +1,90 @@
|
||||
/*
|
||||
* Copyright (c) 2004-2005 The Trustees of Indiana University and Indiana
|
||||
* University Research and Technology
|
||||
* Corporation. All rights reserved.
|
||||
* Copyright (c) 2004-2006 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.
|
||||
* $COPYRIGHT$
|
||||
*
|
||||
* Additional copyrights may follow
|
||||
*
|
||||
* $HEADER$
|
||||
*/
|
||||
/**
|
||||
* @file:
|
||||
* Part of the process launcher. See pls_process.h for an overview of how it works.
|
||||
*/
|
||||
|
||||
#ifndef ORTE_PLS_RSH_EXPORT_H
|
||||
#define ORTE_PLS_RSH_EXPORT_H
|
||||
|
||||
#include "orte_config.h"
|
||||
|
||||
#include "opal/threads/condition.h"
|
||||
#include "opal/mca/mca.h"
|
||||
#include "orte/mca/pls/pls.h"
|
||||
|
||||
#if defined(c_plusplus) || defined(__cplusplus)
|
||||
extern "C" {
|
||||
#endif
|
||||
|
||||
/*
|
||||
* Module open / close
|
||||
*/
|
||||
int orte_pls_process_component_open(void);
|
||||
int orte_pls_process_component_close(void);
|
||||
orte_pls_base_module_t* orte_pls_process_component_init(int *priority);
|
||||
|
||||
/*
|
||||
* Startup / Shutdown
|
||||
*/
|
||||
int orte_pls_process_finalize(void);
|
||||
|
||||
/*
|
||||
* Interface
|
||||
*/
|
||||
int orte_pls_process_launch(orte_jobid_t);
|
||||
int orte_pls_process_terminate_job(orte_jobid_t, opal_list_t*);
|
||||
int orte_pls_process_terminate_orteds(orte_jobid_t, opal_list_t*);
|
||||
int orte_pls_process_terminate_proc(const orte_process_name_t* proc_name);
|
||||
int orte_pls_process_signal_job(orte_jobid_t, int32_t, opal_list_t*);
|
||||
int orte_pls_process_signal_proc(const orte_process_name_t* proc_name, int32_t);
|
||||
|
||||
/**
|
||||
* PLS Component
|
||||
*/
|
||||
struct orte_pls_process_component_t {
|
||||
orte_pls_base_component_t super;
|
||||
bool debug;
|
||||
bool debug_malloc;
|
||||
bool debug_daemons;
|
||||
bool timing;
|
||||
bool reap;
|
||||
bool assume_same_shell;
|
||||
bool force_process;
|
||||
int delay;
|
||||
int priority;
|
||||
char *agent_param;
|
||||
char** agent_argv;
|
||||
int agent_argc;
|
||||
char* agent_path;
|
||||
char* orted;
|
||||
orte_std_cntr_t num_children;
|
||||
orte_std_cntr_t num_concurrent;
|
||||
opal_mutex_t lock;
|
||||
opal_condition_t cond;
|
||||
};
|
||||
typedef struct orte_pls_process_component_t orte_pls_process_component_t;
|
||||
|
||||
ORTE_MODULE_DECLSPEC extern orte_pls_process_component_t mca_pls_process_component;
|
||||
extern orte_pls_base_module_t orte_pls_process_module;
|
||||
|
||||
#if defined(c_plusplus) || defined(__cplusplus)
|
||||
}
|
||||
#endif
|
||||
#endif /* ORTE_PLS_RSH_EXPORT_H */
|
322
orte/mca/pls/process/pls_process_component.c
Обычный файл
322
orte/mca/pls/process/pls_process_component.c
Обычный файл
@ -0,0 +1,322 @@
|
||||
/*
|
||||
* Copyright (c) 2004-2005 The Trustees of Indiana University and Indiana
|
||||
* University Research and Technology
|
||||
* Corporation. All rights reserved.
|
||||
* Copyright (c) 2004-2006 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.
|
||||
* $COPYRIGHT$
|
||||
*
|
||||
* Additional copyrights may follow
|
||||
*
|
||||
* $HEADER$
|
||||
*
|
||||
* These symbols are in a file by themselves to provide nice linker
|
||||
* semantics. Since linkers generally pull in symbols by object
|
||||
* files, keeping these symbols as the only symbols in this file
|
||||
* prevents utility programs such as "ompi_info" from having to import
|
||||
* entire components just to query their version and parameters.
|
||||
*/
|
||||
|
||||
#include "orte_config.h"
|
||||
#include "orte/orte_constants.h"
|
||||
|
||||
#include <stdlib.h>
|
||||
#ifdef HAVE_UNISTD_H
|
||||
#include <unistd.h>
|
||||
#endif
|
||||
#include <ctype.h>
|
||||
|
||||
#include "opal/util/argv.h"
|
||||
#include "opal/util/path.h"
|
||||
#include "opal/util/basename.h"
|
||||
#include "opal/util/show_help.h"
|
||||
#include "opal/mca/base/mca_base_param.h"
|
||||
|
||||
#include "orte/mca/errmgr/errmgr.h"
|
||||
#include "orte/mca/rml/rml.h"
|
||||
|
||||
#include "orte/mca/pls/pls.h"
|
||||
#include "orte/mca/pls/base/pls_private.h"
|
||||
#include "orte/mca/pls/process/pls_process.h"
|
||||
|
||||
#if !defined(__WINDOWS__)
|
||||
extern char **environ;
|
||||
#endif /* !defined(__WINDOWS__) */
|
||||
|
||||
/*
|
||||
* Local function
|
||||
*/
|
||||
static char **search(const char* agent_list);
|
||||
|
||||
|
||||
/*
|
||||
* Public string showing the pls ompi_process component version number
|
||||
*/
|
||||
const char *mca_pls_process_component_version_string =
|
||||
"Open MPI process pls MCA component version " ORTE_VERSION;
|
||||
|
||||
|
||||
/*
|
||||
* Instantiate the public struct with all of our public information
|
||||
* and pointers to our public functions in it
|
||||
*/
|
||||
|
||||
orte_pls_process_component_t mca_pls_process_component = {
|
||||
{
|
||||
/* First, the mca_component_t struct containing meta information
|
||||
about the component itself */
|
||||
|
||||
{
|
||||
/* Indicate that we are a pls v1.3.0 component (which also
|
||||
implies a specific MCA version) */
|
||||
|
||||
ORTE_PLS_BASE_VERSION_1_3_0,
|
||||
|
||||
/* Component name and version */
|
||||
|
||||
"process",
|
||||
ORTE_MAJOR_VERSION,
|
||||
ORTE_MINOR_VERSION,
|
||||
ORTE_RELEASE_VERSION,
|
||||
|
||||
/* Component open and close functions */
|
||||
|
||||
orte_pls_process_component_open,
|
||||
orte_pls_process_component_close
|
||||
},
|
||||
|
||||
/* Next the MCA v1.0.0 component meta data */
|
||||
|
||||
{
|
||||
/* Whether the component is checkpointable or not */
|
||||
|
||||
true
|
||||
},
|
||||
|
||||
/* Initialization / querying functions */
|
||||
|
||||
orte_pls_process_component_init
|
||||
}
|
||||
};
|
||||
|
||||
|
||||
|
||||
int orte_pls_process_component_open(void)
|
||||
{
|
||||
int tmp, value;
|
||||
mca_base_component_t *c = &mca_pls_process_component.super.pls_version;
|
||||
|
||||
/* initialize globals */
|
||||
OBJ_CONSTRUCT(&mca_pls_process_component.lock, opal_mutex_t);
|
||||
OBJ_CONSTRUCT(&mca_pls_process_component.cond, opal_condition_t);
|
||||
mca_pls_process_component.num_children = 0;
|
||||
mca_pls_process_component.agent_argv = NULL;
|
||||
mca_pls_process_component.agent_argc = 0;
|
||||
mca_pls_process_component.agent_path = NULL;
|
||||
|
||||
/* lookup parameters */
|
||||
mca_base_param_reg_int(c, "debug",
|
||||
"Whether or not to enable debugging output for the process pls component (0 or 1)",
|
||||
false, false, false, &tmp);
|
||||
mca_pls_process_component.debug = OPAL_INT_TO_BOOL(tmp);
|
||||
mca_base_param_reg_int(c, "num_concurrent",
|
||||
"How many pls_process_agent instances to invoke concurrently (must be > 0)",
|
||||
false, false, 128, &tmp);
|
||||
if (tmp <= 0) {
|
||||
opal_show_help("help-pls-process.txt", "concurrency-less-than-zero",
|
||||
true, tmp);
|
||||
tmp = 1;
|
||||
}
|
||||
mca_pls_process_component.num_concurrent = tmp;
|
||||
|
||||
mca_base_param_reg_int(c, "force_process",
|
||||
"Force the launcher to always use process, even for local daemons",
|
||||
false, false, false, &tmp);
|
||||
mca_pls_process_component.force_process = OPAL_INT_TO_BOOL(tmp);
|
||||
|
||||
if (mca_pls_process_component.debug == 0) {
|
||||
mca_base_param_reg_int_name("orte", "debug",
|
||||
"Whether or not to enable debugging output for all ORTE components (0 or 1)",
|
||||
false, false, false, &tmp);
|
||||
mca_pls_process_component.debug = OPAL_INT_TO_BOOL(tmp);
|
||||
}
|
||||
mca_base_param_reg_int_name("orte", "debug_daemons",
|
||||
"Whether or not to enable debugging of daemons (0 or 1)",
|
||||
false, false, false, &tmp);
|
||||
mca_pls_process_component.debug_daemons = OPAL_INT_TO_BOOL(tmp);
|
||||
|
||||
tmp = mca_base_param_reg_int_name("orte", "timing",
|
||||
"Request that critical timing loops be measured",
|
||||
false, false, 0, &value);
|
||||
if (value != 0) {
|
||||
mca_pls_process_component.timing = true;
|
||||
} else {
|
||||
mca_pls_process_component.timing = false;
|
||||
}
|
||||
|
||||
mca_base_param_reg_string(c, "orted",
|
||||
"The command name that the process pls component will invoke for the ORTE daemon",
|
||||
false, false, "orted",
|
||||
&mca_pls_process_component.orted);
|
||||
|
||||
mca_base_param_reg_int(c, "priority",
|
||||
"Priority of the process pls component",
|
||||
false, false, 10,
|
||||
&mca_pls_process_component.priority);
|
||||
mca_base_param_reg_int(c, "delay",
|
||||
"Delay (in seconds) between invocations of the remote agent, but only used when the \"debug\" MCA parameter is true, or the top-level MCA debugging is enabled (otherwise this value is ignored)",
|
||||
false, false, 1,
|
||||
&mca_pls_process_component.delay);
|
||||
mca_base_param_reg_int(c, "reap",
|
||||
"If set to 1, wait for all the processes to complete before exiting. Otherwise, quit immediately -- without waiting for confirmation that all other processes in the job have completed.",
|
||||
false, false, 1, &tmp);
|
||||
mca_pls_process_component.reap = OPAL_INT_TO_BOOL(tmp);
|
||||
mca_base_param_reg_int(c, "assume_same_shell",
|
||||
"If set to 1, assume that the shell on the remote node is the same as the shell on the local node. Otherwise, probe for what the remote shell.",
|
||||
false, false, 1, &tmp);
|
||||
mca_pls_process_component.assume_same_shell = OPAL_INT_TO_BOOL(tmp);
|
||||
|
||||
mca_base_param_reg_string(c, "agent",
|
||||
"The command used to launch executables on remote nodes (typically either \"ssh\" or \"process\")",
|
||||
false, false, "ssh : process",
|
||||
&mca_pls_process_component.agent_param);
|
||||
|
||||
return ORTE_SUCCESS;
|
||||
}
|
||||
|
||||
|
||||
#if !defined(__WINDOWS__)
|
||||
extern char **environ;
|
||||
#endif /* !defined(__WINDOWS__) */
|
||||
|
||||
orte_pls_base_module_t *orte_pls_process_component_init(int *priority)
|
||||
{
|
||||
char *bname;
|
||||
size_t i;
|
||||
|
||||
/* if we are not an HNP, then don't select us */
|
||||
if (!orte_process_info.seed) {
|
||||
return NULL;
|
||||
}
|
||||
|
||||
/* Take the string that was given to us by the pla_process_agent MCA
|
||||
param and search for it */
|
||||
mca_pls_process_component.agent_argv =
|
||||
search(mca_pls_process_component.agent_param);
|
||||
mca_pls_process_component.agent_argc =
|
||||
opal_argv_count(mca_pls_process_component.agent_argv);
|
||||
mca_pls_process_component.agent_path = NULL;
|
||||
if (mca_pls_process_component.agent_argc > 0) {
|
||||
/* If the agent is ssh, and debug was not selected, then
|
||||
automatically add "-x" */
|
||||
|
||||
bname = opal_basename(mca_pls_process_component.agent_argv[0]);
|
||||
if (NULL != bname && 0 == strcmp(bname, "ssh") &&
|
||||
mca_pls_process_component.debug == 0) {
|
||||
for (i = 1; NULL != mca_pls_process_component.agent_argv[i]; ++i) {
|
||||
if (0 == strcasecmp("-x",
|
||||
mca_pls_process_component.agent_argv[i])) {
|
||||
break;
|
||||
}
|
||||
}
|
||||
if (NULL == mca_pls_process_component.agent_argv[i]) {
|
||||
opal_argv_append(&mca_pls_process_component.agent_argc,
|
||||
&mca_pls_process_component.agent_argv, "-x");
|
||||
}
|
||||
}
|
||||
if (NULL != bname) {
|
||||
free(bname);
|
||||
}
|
||||
}
|
||||
|
||||
/* If we didn't find the agent in the path, then don't use this
|
||||
component */
|
||||
if (NULL == mca_pls_process_component.agent_argv ||
|
||||
NULL == mca_pls_process_component.agent_argv[0]) {
|
||||
return NULL;
|
||||
}
|
||||
mca_pls_process_component.agent_path =
|
||||
opal_path_findv(mca_pls_process_component.agent_argv[0], X_OK,
|
||||
environ, NULL);
|
||||
if (NULL == mca_pls_process_component.agent_path) {
|
||||
return NULL;
|
||||
}
|
||||
*priority = mca_pls_process_component.priority;
|
||||
|
||||
return &orte_pls_process_module;
|
||||
}
|
||||
|
||||
|
||||
int orte_pls_process_component_close(void)
|
||||
{
|
||||
/* cleanup state */
|
||||
OBJ_DESTRUCT(&mca_pls_process_component.lock);
|
||||
OBJ_DESTRUCT(&mca_pls_process_component.cond);
|
||||
if (NULL != mca_pls_process_component.orted) {
|
||||
free(mca_pls_process_component.orted);
|
||||
}
|
||||
if (NULL != mca_pls_process_component.agent_param) {
|
||||
free(mca_pls_process_component.agent_param);
|
||||
}
|
||||
if (NULL != mca_pls_process_component.agent_argv) {
|
||||
opal_argv_free(mca_pls_process_component.agent_argv);
|
||||
}
|
||||
if (NULL != mca_pls_process_component.agent_path) {
|
||||
free(mca_pls_process_component.agent_path);
|
||||
}
|
||||
return ORTE_SUCCESS;
|
||||
}
|
||||
|
||||
|
||||
/*
|
||||
* Take a colon-delimited list of agents and locate the first one that
|
||||
* we are able to find in the PATH. Split that one into argv and
|
||||
* return it. If nothing found, then return NULL.
|
||||
*/
|
||||
static char **search(const char* agent_list)
|
||||
{
|
||||
int i, j;
|
||||
char *line, **lines = opal_argv_split(agent_list, ':');
|
||||
char **tokens, *tmp;
|
||||
char cwd[PATH_MAX];
|
||||
|
||||
getcwd(cwd, PATH_MAX);
|
||||
for (i = 0; NULL != lines[i]; ++i) {
|
||||
line = lines[i];
|
||||
|
||||
/* Trim whitespace at the beginning and end of the line */
|
||||
for (j = 0; '\0' != line[j] && isspace(line[j]); ++line) {
|
||||
continue;
|
||||
}
|
||||
for (j = strlen(line) - 2; j > 0 && isspace(line[j]); ++j) {
|
||||
line[j] = '\0';
|
||||
}
|
||||
if (strlen(line) <= 0) {
|
||||
continue;
|
||||
}
|
||||
|
||||
/* Split it */
|
||||
tokens = opal_argv_split(line, ' ');
|
||||
|
||||
/* Look for the first token in the PATH */
|
||||
tmp = opal_path_findv(tokens[0], X_OK, environ, cwd);
|
||||
if (NULL != tmp) {
|
||||
free(tokens[0]);
|
||||
tokens[0] = tmp;
|
||||
opal_argv_free(lines);
|
||||
return tokens;
|
||||
}
|
||||
|
||||
/* Didn't find it */
|
||||
opal_argv_free(tokens);
|
||||
}
|
||||
|
||||
/* Doh -- didn't find anything */
|
||||
opal_argv_free(lines);
|
||||
return NULL;
|
||||
}
|
1284
orte/mca/pls/process/pls_process_module.c
Обычный файл
1284
orte/mca/pls/process/pls_process_module.c
Обычный файл
Разница между файлами не показана из-за своего большого размера
Загрузить разницу
Загрузка…
Ссылка в новой задаче
Block a user