1
1
openmpi/orte/mca/rmaps/base/rmaps_base_node.c

390 строки
13 KiB
C
Исходник Обычный вид История

This patch fixes several issues relating to comm_spawn and N1GE. In particular, it does the following: 1. Modifies the RAS framework so it correctly stores and retrieves the actual slots in use, not just those that were allocated. Although the RAS node structure had storage for the number of slots in use, it turned out that the base function for storing and retrieving that information ignored what was in the field and simply set it equal to the number of slots allocated. This has now been fixed. 2. Modified the RMAPS framework so it updates the registry with the actual number of slots used by the mapping. Note that daemons are still NOT counted in this process as daemons are NOT mapped at this time. This will be fixed in 2.0, but will not be addressed in 1.x. 3. Added a new MCA parameter "rmaps_base_no_oversubscribe" that tells the system not to oversubscribe nodes even if the underlying environment permits it. The default is to oversubscribe if needed and the underlying environment permits it. I'm sure someone may argue "why would a user do that?", but it turns out that (looking ahead to dynamic resource reservations) sometimes users won't know how many nodes or slots they've been given in advance - this just allows them to say "hey, I'd rather not run if I didn't get enough". 4. Reorganizes the RMAPS framework to more easily support multiple components. A lot of the logic in the round_robin mapper was very valuable to any component - this has been moved to the base so others can take advantage of it. 5. Added a new test program "hello_nodename" - just does "hello_world" but also prints out the name of the node it is on. 6. Made the orte_ras_node_t object a full ORTE data type so it can more easily be copied, packed, etc. This proved helpful for the RMAPS code reorganization and might be of use elsewhere too. This commit was SVN r10697.
2006-07-10 14:10:21 +00:00
/*
* Copyright (c) 2004-2005 The Trustees of Indiana University and Indiana
* University Research and Technology
* Corporation. All rights reserved.
* Copyright (c) 2004-2005 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
*
* $HEADER$
*/
#include "orte_config.h"
#include "orte/orte_constants.h"
#include <string.h>
#include "opal/util/output.h"
#include "opal/util/argv.h"
#include "opal/mca/mca.h"
#include "opal/mca/base/base.h"
#include "opal/mca/base/mca_base_param.h"
#include "opal/util/if.h"
#include "opal/util/show_help.h"
#include "orte/util/sys_info.h"
#include "orte/mca/errmgr/errmgr.h"
#include "orte/mca/soh/soh_types.h"
#include "orte/mca/ras/base/ras_base_node.h"
#include "orte/mca/gpr/gpr.h"
#include "orte/mca/ns/ns.h"
#include "orte/mca/rmaps/base/base.h"
#include "orte/mca/rmaps/base/rmaps_base_map.h"
#include "orte/mca/rmaps/base/rmaps_base_node.h"
/*
* A sanity check to ensure that all of the requested nodes are actually
* allocated to this application.
*/
static bool are_all_mapped_valid(char **mapping,
int num_mapped,
opal_list_t* nodes)
{
opal_list_item_t *item;
int i;
bool matched;
for (i = 0; i < num_mapped; ++i) {
matched = false;
for(item = opal_list_get_first(nodes);
item != opal_list_get_end(nodes);
item = opal_list_get_next(item) ) {
if( 0 == strcmp( ((orte_ras_node_t*) item)->node_name, mapping[i]) ) {
matched = true;
break;
}
}
/* If we find one requested resource that is not allocated,
* then return an error */
if(!matched) {
return false;
}
}
return true;
}
/*
* If the node in question is in the current mapping.
*/
static bool is_mapped(opal_list_item_t *item,
char **mapping,
int num_mapped)
{
int i;
for ( i = 0; i < num_mapped; ++i) {
if ( 0 == strcmp( ((orte_ras_node_t*) item)->node_name, mapping[i])){
return true;
}
}
return false;
}
/*
* Query the registry for all nodes allocated to a specified job
*/
int orte_rmaps_base_get_target_nodes(opal_list_t* nodes, orte_jobid_t jobid, size_t *total_num_slots)
This patch fixes several issues relating to comm_spawn and N1GE. In particular, it does the following: 1. Modifies the RAS framework so it correctly stores and retrieves the actual slots in use, not just those that were allocated. Although the RAS node structure had storage for the number of slots in use, it turned out that the base function for storing and retrieving that information ignored what was in the field and simply set it equal to the number of slots allocated. This has now been fixed. 2. Modified the RMAPS framework so it updates the registry with the actual number of slots used by the mapping. Note that daemons are still NOT counted in this process as daemons are NOT mapped at this time. This will be fixed in 2.0, but will not be addressed in 1.x. 3. Added a new MCA parameter "rmaps_base_no_oversubscribe" that tells the system not to oversubscribe nodes even if the underlying environment permits it. The default is to oversubscribe if needed and the underlying environment permits it. I'm sure someone may argue "why would a user do that?", but it turns out that (looking ahead to dynamic resource reservations) sometimes users won't know how many nodes or slots they've been given in advance - this just allows them to say "hey, I'd rather not run if I didn't get enough". 4. Reorganizes the RMAPS framework to more easily support multiple components. A lot of the logic in the round_robin mapper was very valuable to any component - this has been moved to the base so others can take advantage of it. 5. Added a new test program "hello_nodename" - just does "hello_world" but also prints out the name of the node it is on. 6. Made the orte_ras_node_t object a full ORTE data type so it can more easily be copied, packed, etc. This proved helpful for the RMAPS code reorganization and might be of use elsewhere too. This commit was SVN r10697.
2006-07-10 14:10:21 +00:00
{
opal_list_item_t *item, *next;
orte_ras_node_t *node;
int id, rc, nolocal;
size_t num_slots=0;
This patch fixes several issues relating to comm_spawn and N1GE. In particular, it does the following: 1. Modifies the RAS framework so it correctly stores and retrieves the actual slots in use, not just those that were allocated. Although the RAS node structure had storage for the number of slots in use, it turned out that the base function for storing and retrieving that information ignored what was in the field and simply set it equal to the number of slots allocated. This has now been fixed. 2. Modified the RMAPS framework so it updates the registry with the actual number of slots used by the mapping. Note that daemons are still NOT counted in this process as daemons are NOT mapped at this time. This will be fixed in 2.0, but will not be addressed in 1.x. 3. Added a new MCA parameter "rmaps_base_no_oversubscribe" that tells the system not to oversubscribe nodes even if the underlying environment permits it. The default is to oversubscribe if needed and the underlying environment permits it. I'm sure someone may argue "why would a user do that?", but it turns out that (looking ahead to dynamic resource reservations) sometimes users won't know how many nodes or slots they've been given in advance - this just allows them to say "hey, I'd rather not run if I didn't get enough". 4. Reorganizes the RMAPS framework to more easily support multiple components. A lot of the logic in the round_robin mapper was very valuable to any component - this has been moved to the base so others can take advantage of it. 5. Added a new test program "hello_nodename" - just does "hello_world" but also prints out the name of the node it is on. 6. Made the orte_ras_node_t object a full ORTE data type so it can more easily be copied, packed, etc. This proved helpful for the RMAPS code reorganization and might be of use elsewhere too. This commit was SVN r10697.
2006-07-10 14:10:21 +00:00
/** set default answer */
*total_num_slots = 0;
This patch fixes several issues relating to comm_spawn and N1GE. In particular, it does the following: 1. Modifies the RAS framework so it correctly stores and retrieves the actual slots in use, not just those that were allocated. Although the RAS node structure had storage for the number of slots in use, it turned out that the base function for storing and retrieving that information ignored what was in the field and simply set it equal to the number of slots allocated. This has now been fixed. 2. Modified the RMAPS framework so it updates the registry with the actual number of slots used by the mapping. Note that daemons are still NOT counted in this process as daemons are NOT mapped at this time. This will be fixed in 2.0, but will not be addressed in 1.x. 3. Added a new MCA parameter "rmaps_base_no_oversubscribe" that tells the system not to oversubscribe nodes even if the underlying environment permits it. The default is to oversubscribe if needed and the underlying environment permits it. I'm sure someone may argue "why would a user do that?", but it turns out that (looking ahead to dynamic resource reservations) sometimes users won't know how many nodes or slots they've been given in advance - this just allows them to say "hey, I'd rather not run if I didn't get enough". 4. Reorganizes the RMAPS framework to more easily support multiple components. A lot of the logic in the round_robin mapper was very valuable to any component - this has been moved to the base so others can take advantage of it. 5. Added a new test program "hello_nodename" - just does "hello_world" but also prints out the name of the node it is on. 6. Made the orte_ras_node_t object a full ORTE data type so it can more easily be copied, packed, etc. This proved helpful for the RMAPS code reorganization and might be of use elsewhere too. This commit was SVN r10697.
2006-07-10 14:10:21 +00:00
if(ORTE_SUCCESS != (rc = orte_ras_base_node_query_alloc(nodes, jobid))) {
ORTE_ERROR_LOG(rc);
return rc;
}
/* If the "no local" option was set, then remove the local node
from the list */
id = mca_base_param_find("rmaps", NULL, "base_schedule_local");
mca_base_param_lookup_int(id, &nolocal);
if (0 == nolocal) {
for (item = opal_list_get_first(nodes);
item != opal_list_get_end(nodes);
item = opal_list_get_next(item) ) {
if (0 == strcmp(((orte_ras_node_t *) item)->node_name,
orte_system_info.nodename) ||
opal_ifislocal(((orte_ras_node_t *) item)->node_name)) {
opal_list_remove_item(nodes, item);
break;
}
}
}
/** remove all nodes that are already at max usage */
item = opal_list_get_first(nodes);
while (item != opal_list_get_end(nodes)) {
/** save the next pointer in case we remove this node */
next = opal_list_get_next(item);
/** check to see if this node is fully used - remove if so */
node = (orte_ras_node_t*)item;
if (0 != node->node_slots_max && node->node_slots_inuse > node->node_slots_max) {
opal_list_remove_item(nodes, item);
} else { /** otherwise, add its slots to the total */
num_slots += node->node_slots;
This patch fixes several issues relating to comm_spawn and N1GE. In particular, it does the following: 1. Modifies the RAS framework so it correctly stores and retrieves the actual slots in use, not just those that were allocated. Although the RAS node structure had storage for the number of slots in use, it turned out that the base function for storing and retrieving that information ignored what was in the field and simply set it equal to the number of slots allocated. This has now been fixed. 2. Modified the RMAPS framework so it updates the registry with the actual number of slots used by the mapping. Note that daemons are still NOT counted in this process as daemons are NOT mapped at this time. This will be fixed in 2.0, but will not be addressed in 1.x. 3. Added a new MCA parameter "rmaps_base_no_oversubscribe" that tells the system not to oversubscribe nodes even if the underlying environment permits it. The default is to oversubscribe if needed and the underlying environment permits it. I'm sure someone may argue "why would a user do that?", but it turns out that (looking ahead to dynamic resource reservations) sometimes users won't know how many nodes or slots they've been given in advance - this just allows them to say "hey, I'd rather not run if I didn't get enough". 4. Reorganizes the RMAPS framework to more easily support multiple components. A lot of the logic in the round_robin mapper was very valuable to any component - this has been moved to the base so others can take advantage of it. 5. Added a new test program "hello_nodename" - just does "hello_world" but also prints out the name of the node it is on. 6. Made the orte_ras_node_t object a full ORTE data type so it can more easily be copied, packed, etc. This proved helpful for the RMAPS code reorganization and might be of use elsewhere too. This commit was SVN r10697.
2006-07-10 14:10:21 +00:00
}
/** go on to next item */
item = next;
}
/* Sanity check to make sure we have been allocated nodes */
if (0 == opal_list_get_size(nodes)) {
ORTE_ERROR_LOG(ORTE_ERR_TEMP_OUT_OF_RESOURCE);
return ORTE_ERR_TEMP_OUT_OF_RESOURCE;
}
*total_num_slots = num_slots;
This patch fixes several issues relating to comm_spawn and N1GE. In particular, it does the following: 1. Modifies the RAS framework so it correctly stores and retrieves the actual slots in use, not just those that were allocated. Although the RAS node structure had storage for the number of slots in use, it turned out that the base function for storing and retrieving that information ignored what was in the field and simply set it equal to the number of slots allocated. This has now been fixed. 2. Modified the RMAPS framework so it updates the registry with the actual number of slots used by the mapping. Note that daemons are still NOT counted in this process as daemons are NOT mapped at this time. This will be fixed in 2.0, but will not be addressed in 1.x. 3. Added a new MCA parameter "rmaps_base_no_oversubscribe" that tells the system not to oversubscribe nodes even if the underlying environment permits it. The default is to oversubscribe if needed and the underlying environment permits it. I'm sure someone may argue "why would a user do that?", but it turns out that (looking ahead to dynamic resource reservations) sometimes users won't know how many nodes or slots they've been given in advance - this just allows them to say "hey, I'd rather not run if I didn't get enough". 4. Reorganizes the RMAPS framework to more easily support multiple components. A lot of the logic in the round_robin mapper was very valuable to any component - this has been moved to the base so others can take advantage of it. 5. Added a new test program "hello_nodename" - just does "hello_world" but also prints out the name of the node it is on. 6. Made the orte_ras_node_t object a full ORTE data type so it can more easily be copied, packed, etc. This proved helpful for the RMAPS code reorganization and might be of use elsewhere too. This commit was SVN r10697.
2006-07-10 14:10:21 +00:00
return ORTE_SUCCESS;
}
/*
* Create a sub-list of nodes to be used for user-specified mappings
*/
int orte_rmaps_base_get_mapped_targets(opal_list_t *mapped_node_list,
orte_app_context_t *app,
opal_list_t *master_node_list,
size_t *total_num_slots)
This patch fixes several issues relating to comm_spawn and N1GE. In particular, it does the following: 1. Modifies the RAS framework so it correctly stores and retrieves the actual slots in use, not just those that were allocated. Although the RAS node structure had storage for the number of slots in use, it turned out that the base function for storing and retrieving that information ignored what was in the field and simply set it equal to the number of slots allocated. This has now been fixed. 2. Modified the RMAPS framework so it updates the registry with the actual number of slots used by the mapping. Note that daemons are still NOT counted in this process as daemons are NOT mapped at this time. This will be fixed in 2.0, but will not be addressed in 1.x. 3. Added a new MCA parameter "rmaps_base_no_oversubscribe" that tells the system not to oversubscribe nodes even if the underlying environment permits it. The default is to oversubscribe if needed and the underlying environment permits it. I'm sure someone may argue "why would a user do that?", but it turns out that (looking ahead to dynamic resource reservations) sometimes users won't know how many nodes or slots they've been given in advance - this just allows them to say "hey, I'd rather not run if I didn't get enough". 4. Reorganizes the RMAPS framework to more easily support multiple components. A lot of the logic in the round_robin mapper was very valuable to any component - this has been moved to the base so others can take advantage of it. 5. Added a new test program "hello_nodename" - just does "hello_world" but also prints out the name of the node it is on. 6. Made the orte_ras_node_t object a full ORTE data type so it can more easily be copied, packed, etc. This proved helpful for the RMAPS code reorganization and might be of use elsewhere too. This commit was SVN r10697.
2006-07-10 14:10:21 +00:00
{
orte_app_context_map_t** loc_map = app->map_data;
opal_list_item_t *item;
orte_ras_node_t *node, *new_node;
char **mapped_nodes = NULL;
int num_mapped_nodes = 0;
size_t j, k, num_slots=0;
This patch fixes several issues relating to comm_spawn and N1GE. In particular, it does the following: 1. Modifies the RAS framework so it correctly stores and retrieves the actual slots in use, not just those that were allocated. Although the RAS node structure had storage for the number of slots in use, it turned out that the base function for storing and retrieving that information ignored what was in the field and simply set it equal to the number of slots allocated. This has now been fixed. 2. Modified the RMAPS framework so it updates the registry with the actual number of slots used by the mapping. Note that daemons are still NOT counted in this process as daemons are NOT mapped at this time. This will be fixed in 2.0, but will not be addressed in 1.x. 3. Added a new MCA parameter "rmaps_base_no_oversubscribe" that tells the system not to oversubscribe nodes even if the underlying environment permits it. The default is to oversubscribe if needed and the underlying environment permits it. I'm sure someone may argue "why would a user do that?", but it turns out that (looking ahead to dynamic resource reservations) sometimes users won't know how many nodes or slots they've been given in advance - this just allows them to say "hey, I'd rather not run if I didn't get enough". 4. Reorganizes the RMAPS framework to more easily support multiple components. A lot of the logic in the round_robin mapper was very valuable to any component - this has been moved to the base so others can take advantage of it. 5. Added a new test program "hello_nodename" - just does "hello_world" but also prints out the name of the node it is on. 6. Made the orte_ras_node_t object a full ORTE data type so it can more easily be copied, packed, etc. This proved helpful for the RMAPS code reorganization and might be of use elsewhere too. This commit was SVN r10697.
2006-07-10 14:10:21 +00:00
int rc;
/** set default answer */
*total_num_slots = 0;
This patch fixes several issues relating to comm_spawn and N1GE. In particular, it does the following: 1. Modifies the RAS framework so it correctly stores and retrieves the actual slots in use, not just those that were allocated. Although the RAS node structure had storage for the number of slots in use, it turned out that the base function for storing and retrieving that information ignored what was in the field and simply set it equal to the number of slots allocated. This has now been fixed. 2. Modified the RMAPS framework so it updates the registry with the actual number of slots used by the mapping. Note that daemons are still NOT counted in this process as daemons are NOT mapped at this time. This will be fixed in 2.0, but will not be addressed in 1.x. 3. Added a new MCA parameter "rmaps_base_no_oversubscribe" that tells the system not to oversubscribe nodes even if the underlying environment permits it. The default is to oversubscribe if needed and the underlying environment permits it. I'm sure someone may argue "why would a user do that?", but it turns out that (looking ahead to dynamic resource reservations) sometimes users won't know how many nodes or slots they've been given in advance - this just allows them to say "hey, I'd rather not run if I didn't get enough". 4. Reorganizes the RMAPS framework to more easily support multiple components. A lot of the logic in the round_robin mapper was very valuable to any component - this has been moved to the base so others can take advantage of it. 5. Added a new test program "hello_nodename" - just does "hello_world" but also prints out the name of the node it is on. 6. Made the orte_ras_node_t object a full ORTE data type so it can more easily be copied, packed, etc. This proved helpful for the RMAPS code reorganization and might be of use elsewhere too. This commit was SVN r10697.
2006-07-10 14:10:21 +00:00
/* Accumulate all of the host name mappings */
for(k = 0; k < app->num_map; ++k) {
if ( ORTE_APP_CONTEXT_MAP_HOSTNAME == loc_map[k]->map_type ) {
if(mapped_nodes == NULL) {
mapped_nodes = opal_argv_split(loc_map[k]->map_data, ',');
num_mapped_nodes = opal_argv_count(mapped_nodes);
}
else { /* Append to the existing mapping */
char ** mini_map = opal_argv_split(loc_map[k]->map_data, ',');
size_t mini_num_map = opal_argv_count(mini_map);
for (j = 0; j < mini_num_map; ++j) {
rc = opal_argv_append(&num_mapped_nodes, &mapped_nodes, mini_map[j]);
if (OPAL_SUCCESS != rc) {
ORTE_ERROR_LOG(rc);
return rc;
}
}
opal_argv_free(mini_map);
}
}
}
/** check to see that all the nodes in the specified mapping have been allocated
* for our use - if not, then that's an error
*/
if( !are_all_mapped_valid(mapped_nodes, num_mapped_nodes, master_node_list) ) {
opal_show_help("help-orte-rmaps-base.txt", "orte-rmaps-base:not-all-mapped-alloc",
This patch fixes several issues relating to comm_spawn and N1GE. In particular, it does the following: 1. Modifies the RAS framework so it correctly stores and retrieves the actual slots in use, not just those that were allocated. Although the RAS node structure had storage for the number of slots in use, it turned out that the base function for storing and retrieving that information ignored what was in the field and simply set it equal to the number of slots allocated. This has now been fixed. 2. Modified the RMAPS framework so it updates the registry with the actual number of slots used by the mapping. Note that daemons are still NOT counted in this process as daemons are NOT mapped at this time. This will be fixed in 2.0, but will not be addressed in 1.x. 3. Added a new MCA parameter "rmaps_base_no_oversubscribe" that tells the system not to oversubscribe nodes even if the underlying environment permits it. The default is to oversubscribe if needed and the underlying environment permits it. I'm sure someone may argue "why would a user do that?", but it turns out that (looking ahead to dynamic resource reservations) sometimes users won't know how many nodes or slots they've been given in advance - this just allows them to say "hey, I'd rather not run if I didn't get enough". 4. Reorganizes the RMAPS framework to more easily support multiple components. A lot of the logic in the round_robin mapper was very valuable to any component - this has been moved to the base so others can take advantage of it. 5. Added a new test program "hello_nodename" - just does "hello_world" but also prints out the name of the node it is on. 6. Made the orte_ras_node_t object a full ORTE data type so it can more easily be copied, packed, etc. This proved helpful for the RMAPS code reorganization and might be of use elsewhere too. This commit was SVN r10697.
2006-07-10 14:10:21 +00:00
true, app->app, opal_argv_join(mapped_nodes, ','));
ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
return ORTE_ERR_OUT_OF_RESOURCE;
}
/** setup the working node list to include only those nodes that were specified
* in this mapping. We don't need to worry about nodes being fully used or not
* since the master list only includes nodes that aren't.
*/
for (item = opal_list_get_first(master_node_list);
item != opal_list_get_end(master_node_list);
item = opal_list_get_next(item) ) {
node = (orte_ras_node_t*)item;
if( is_mapped(item, mapped_nodes, num_mapped_nodes) ) {
/** we can't just add this item to the mapped_node_list as it cannot be
* on two lists at the same time, so we need to copy it first
*/
if (ORTE_SUCCESS != (rc = orte_dss.copy((void**)&new_node, node, ORTE_RAS_NODE))) {
This patch fixes several issues relating to comm_spawn and N1GE. In particular, it does the following: 1. Modifies the RAS framework so it correctly stores and retrieves the actual slots in use, not just those that were allocated. Although the RAS node structure had storage for the number of slots in use, it turned out that the base function for storing and retrieving that information ignored what was in the field and simply set it equal to the number of slots allocated. This has now been fixed. 2. Modified the RMAPS framework so it updates the registry with the actual number of slots used by the mapping. Note that daemons are still NOT counted in this process as daemons are NOT mapped at this time. This will be fixed in 2.0, but will not be addressed in 1.x. 3. Added a new MCA parameter "rmaps_base_no_oversubscribe" that tells the system not to oversubscribe nodes even if the underlying environment permits it. The default is to oversubscribe if needed and the underlying environment permits it. I'm sure someone may argue "why would a user do that?", but it turns out that (looking ahead to dynamic resource reservations) sometimes users won't know how many nodes or slots they've been given in advance - this just allows them to say "hey, I'd rather not run if I didn't get enough". 4. Reorganizes the RMAPS framework to more easily support multiple components. A lot of the logic in the round_robin mapper was very valuable to any component - this has been moved to the base so others can take advantage of it. 5. Added a new test program "hello_nodename" - just does "hello_world" but also prints out the name of the node it is on. 6. Made the orte_ras_node_t object a full ORTE data type so it can more easily be copied, packed, etc. This proved helpful for the RMAPS code reorganization and might be of use elsewhere too. This commit was SVN r10697.
2006-07-10 14:10:21 +00:00
ORTE_ERROR_LOG(rc);
return rc;
}
opal_list_append(mapped_node_list, &new_node->super);
num_slots += new_node->node_slots;
This patch fixes several issues relating to comm_spawn and N1GE. In particular, it does the following: 1. Modifies the RAS framework so it correctly stores and retrieves the actual slots in use, not just those that were allocated. Although the RAS node structure had storage for the number of slots in use, it turned out that the base function for storing and retrieving that information ignored what was in the field and simply set it equal to the number of slots allocated. This has now been fixed. 2. Modified the RMAPS framework so it updates the registry with the actual number of slots used by the mapping. Note that daemons are still NOT counted in this process as daemons are NOT mapped at this time. This will be fixed in 2.0, but will not be addressed in 1.x. 3. Added a new MCA parameter "rmaps_base_no_oversubscribe" that tells the system not to oversubscribe nodes even if the underlying environment permits it. The default is to oversubscribe if needed and the underlying environment permits it. I'm sure someone may argue "why would a user do that?", but it turns out that (looking ahead to dynamic resource reservations) sometimes users won't know how many nodes or slots they've been given in advance - this just allows them to say "hey, I'd rather not run if I didn't get enough". 4. Reorganizes the RMAPS framework to more easily support multiple components. A lot of the logic in the round_robin mapper was very valuable to any component - this has been moved to the base so others can take advantage of it. 5. Added a new test program "hello_nodename" - just does "hello_world" but also prints out the name of the node it is on. 6. Made the orte_ras_node_t object a full ORTE data type so it can more easily be copied, packed, etc. This proved helpful for the RMAPS code reorganization and might be of use elsewhere too. This commit was SVN r10697.
2006-07-10 14:10:21 +00:00
}
}
/** check that anything is left! */
if (0 == opal_list_get_size(mapped_node_list)) {
opal_show_help("help-orte-rmaps-base.txt", "orte-rmaps-base:no-mapped-node",
This patch fixes several issues relating to comm_spawn and N1GE. In particular, it does the following: 1. Modifies the RAS framework so it correctly stores and retrieves the actual slots in use, not just those that were allocated. Although the RAS node structure had storage for the number of slots in use, it turned out that the base function for storing and retrieving that information ignored what was in the field and simply set it equal to the number of slots allocated. This has now been fixed. 2. Modified the RMAPS framework so it updates the registry with the actual number of slots used by the mapping. Note that daemons are still NOT counted in this process as daemons are NOT mapped at this time. This will be fixed in 2.0, but will not be addressed in 1.x. 3. Added a new MCA parameter "rmaps_base_no_oversubscribe" that tells the system not to oversubscribe nodes even if the underlying environment permits it. The default is to oversubscribe if needed and the underlying environment permits it. I'm sure someone may argue "why would a user do that?", but it turns out that (looking ahead to dynamic resource reservations) sometimes users won't know how many nodes or slots they've been given in advance - this just allows them to say "hey, I'd rather not run if I didn't get enough". 4. Reorganizes the RMAPS framework to more easily support multiple components. A lot of the logic in the round_robin mapper was very valuable to any component - this has been moved to the base so others can take advantage of it. 5. Added a new test program "hello_nodename" - just does "hello_world" but also prints out the name of the node it is on. 6. Made the orte_ras_node_t object a full ORTE data type so it can more easily be copied, packed, etc. This proved helpful for the RMAPS code reorganization and might be of use elsewhere too. This commit was SVN r10697.
2006-07-10 14:10:21 +00:00
true, app->num_procs, app->app);
ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
return ORTE_ERR_OUT_OF_RESOURCE;
}
*total_num_slots = num_slots;
This patch fixes several issues relating to comm_spawn and N1GE. In particular, it does the following: 1. Modifies the RAS framework so it correctly stores and retrieves the actual slots in use, not just those that were allocated. Although the RAS node structure had storage for the number of slots in use, it turned out that the base function for storing and retrieving that information ignored what was in the field and simply set it equal to the number of slots allocated. This has now been fixed. 2. Modified the RMAPS framework so it updates the registry with the actual number of slots used by the mapping. Note that daemons are still NOT counted in this process as daemons are NOT mapped at this time. This will be fixed in 2.0, but will not be addressed in 1.x. 3. Added a new MCA parameter "rmaps_base_no_oversubscribe" that tells the system not to oversubscribe nodes even if the underlying environment permits it. The default is to oversubscribe if needed and the underlying environment permits it. I'm sure someone may argue "why would a user do that?", but it turns out that (looking ahead to dynamic resource reservations) sometimes users won't know how many nodes or slots they've been given in advance - this just allows them to say "hey, I'd rather not run if I didn't get enough". 4. Reorganizes the RMAPS framework to more easily support multiple components. A lot of the logic in the round_robin mapper was very valuable to any component - this has been moved to the base so others can take advantage of it. 5. Added a new test program "hello_nodename" - just does "hello_world" but also prints out the name of the node it is on. 6. Made the orte_ras_node_t object a full ORTE data type so it can more easily be copied, packed, etc. This proved helpful for the RMAPS code reorganization and might be of use elsewhere too. This commit was SVN r10697.
2006-07-10 14:10:21 +00:00
return ORTE_SUCCESS;
}
/*
* Claim a slot for a specified job on a node
*/
int orte_rmaps_base_claim_slot(orte_rmaps_base_map_t *map,
orte_ras_node_t *current_node,
orte_jobid_t jobid, orte_vpid_t vpid,
int proc_index,
opal_list_t *nodes,
opal_list_t *fully_used_nodes)
{
orte_rmaps_base_proc_t *proc;
orte_process_name_t *proc_name;
orte_rmaps_base_node_t *rmaps_node;
int rc;
/* create objects */
rmaps_node = OBJ_NEW(orte_rmaps_base_node_t);
if (NULL == rmaps_node) {
ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
return ORTE_ERR_OUT_OF_RESOURCE;
}
OBJ_RETAIN(current_node);
rmaps_node->node = current_node;
proc = OBJ_NEW(orte_rmaps_base_proc_t);
if (NULL == proc) {
ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
OBJ_RELEASE(rmaps_node);
return ORTE_ERR_OUT_OF_RESOURCE;
}
/* create the process name as an offset from the vpid-start */
rc = orte_ns.create_process_name(&proc_name, current_node->node_cellid,
jobid, vpid);
if (rc != ORTE_SUCCESS) {
ORTE_ERROR_LOG(rc);
OBJ_RELEASE(proc);
OBJ_RELEASE(rmaps_node);
return rc;
}
proc->proc_node = rmaps_node;
proc->proc_name = *proc_name;
proc->proc_rank = vpid;
orte_ns.free_name(&proc_name);
OBJ_RETAIN(proc); /* bump reference count for the node */
opal_list_append(&rmaps_node->node_procs, &proc->super);
map->procs[proc_index] = proc;
/* Save this node on the map */
opal_list_append(&map->nodes, &rmaps_node->super);
/* Be sure to demarcate this slot as claimed for the node */
current_node->node_slots_inuse++;
/* Remove this node if it has reached its max number of allocatable slots OR it has
* reached the soft limit AND we are in a "no oversubscribe" state
*/
if ((0 != current_node->node_slots_max &&
current_node->node_slots_inuse >= current_node->node_slots_max) ||
(!orte_rmaps_base.oversubscribe &&
current_node->node_slots_inuse >= current_node->node_slots)) {
opal_list_remove_item(nodes, (opal_list_item_t*)current_node);
/* add it to the list of fully used nodes */
opal_list_append(fully_used_nodes, &current_node->super);
/** now return the proper code so the caller knows we removed the node! */
return ORTE_ERR_NODE_FULLY_USED;
}
return ORTE_SUCCESS;
}
/*
* Update the node allocations stored in the registry
*/
int orte_rmaps_base_update_node_usage(opal_list_t *nodes)
{
opal_list_item_t* item;
orte_gpr_value_t **values;
int rc;
size_t num_values, i, j;
orte_ras_node_t* node;
num_values = opal_list_get_size(nodes);
if (0 >= num_values) {
ORTE_ERROR_LOG(ORTE_ERR_BAD_PARAM);
return ORTE_ERR_BAD_PARAM;
}
values = (orte_gpr_value_t**)malloc(num_values * sizeof(orte_gpr_value_t*));
if (NULL == values) {
ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
return ORTE_ERR_OUT_OF_RESOURCE;
}
for (i=0; i < num_values; i++) {
if (ORTE_SUCCESS != (rc = orte_gpr.create_value(&(values[i]),
ORTE_GPR_OVERWRITE | ORTE_GPR_TOKENS_AND,
ORTE_NODE_SEGMENT, 1, 0))) {
ORTE_ERROR_LOG(rc);
for (j=0; j < i; j++) {
OBJ_RELEASE(values[j]);
}
free(values);
return rc;
}
}
for(i=0, item = opal_list_get_first(nodes);
i < num_values && item != opal_list_get_end(nodes);
i++, item = opal_list_get_next(item)) {
node = (orte_ras_node_t*)item;
if (ORTE_SUCCESS != (rc = orte_gpr.create_keyval(&(values[i]->keyvals[0]), ORTE_NODE_SLOTS_IN_USE_KEY,
ORTE_SIZE, &(node->node_slots_inuse)))) {
ORTE_ERROR_LOG(rc);
goto cleanup;
}
/* setup index/keys for this node */
rc = orte_schema.get_node_tokens(&(values[i]->tokens), &(values[i]->num_tokens), node->node_cellid, node->node_name);
if (ORTE_SUCCESS != rc) {
ORTE_ERROR_LOG(rc);
goto cleanup;
}
}
/* try the insert */
if (ORTE_SUCCESS != (rc = orte_gpr.put(num_values, values))) {
ORTE_ERROR_LOG(rc);
}
cleanup:
for (j=0; j < num_values; j++) {
OBJ_RELEASE(values[j]);
}
if (NULL != values) free(values);
return rc;
}