22c7f2b3e0
ns_replica.c
- Removed the error logging since I use this function in orte_init_stage1 to
check if we have created a cellid yet or not.
ras_types.h & rase_base_node.h
- This was an empty file. moved the orte_ras_node_t from base/ras_base_node.h
to this file.
- Changed the name of orte_ras_base_node_t to orte_ras_node_t to match the
naming mechanisms in place.
ras.h
- Exposed 2 functions:
- node_insert:
This takes a list of orte_ras_base_node_t's and places them in the Node
Segment of the GPR. This is to be used in orte_init_stage1 for singleton
processes, and the hostfile parsing (see rds_hostfile.c). This just puts
in the appropriate API interface to keep from calling the
orte_ras_base_node_insert function directly.
- node_query:
This is used in hostfile parsing. This just puts in the appropriate API
interface to keep from calling the orte_ras_base_node_query function
directly.
- Touched all of the implemented components to add reference to these new
function pointers
ras_base_select.c & ras_base_open.c
- Add and set the global module reference
rds.h
- Exposed 1 function:
- store_resource:
This stores a list of rds_cell_desc_t's to the Resource Segment.
This is used in conjunction with the orte_ras.node_insert function in
both the orte_init_stage1 for singleton processes and rds_hostfile.c
rds_base_select.c & rds_base_open.c
- Add and set the global module reference
rds_hostfile.c
- Added functionality to create a new cellid for each hostfile, placing
each entry in the hostfile into the same cellid. Currently this is
commented out with the cellid hard coded to 0, with the intention of
taking this out once ORTE is able to handle multiple cellid's
- Instead of just adding hosts to the Node Segment via a direct call to
the ras_base_node_insert() function. First add the hosts to the Resource
Segment of the GPR using the orte_rds.store_resource() function then use
the API version of orte_ras.node_insert() to store the hosts on the Node
Segment.
- Add 1 new function pointer to module as required by the API.
rds_hostfile_component.c
- Converted this to use the new MCA parameter registration
orte_init_stage1.c
- It is possible that a cellid was not created yet for the current environment.
So I put in some logic to test if the cellid 0 existed. If it does then
continue, otherwise create the cellid so we can properly interact with the
GPR via the RDS.
- For the singleton case we insert some 'dummy' data into the GPR. The RAS
matches this logic, so I took out the duplicate GPR put logic, and
replaced it with a call to the orte_ras.node_insert() function.
- Further before calling orte_ras.node_insert() in the singleton case,
we also call orte_rds.store_resource() to add the singleton node to the
Resource Segment.
Console:
- Added a bunch of new functions. Still experimenting with many aspects of the
implementation. This is a checkpoint, and has very limited functionality.
- Should not be considered stable at the moment.
This commit was SVN r6813.
484 строки
14 KiB
C
484 строки
14 KiB
C
/*
|
|
* Copyright (c) 2004-2005 The Trustees of Indiana University.
|
|
* All rights reserved.
|
|
* Copyright (c) 2004-2005 The Trustees of the University of Tennessee.
|
|
* 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 <string.h>
|
|
|
|
#include "include/orte_constants.h"
|
|
#include "opal/util/output.h"
|
|
|
|
#include "mca/errmgr/errmgr.h"
|
|
#include "mca/soh/soh_types.h"
|
|
#include "mca/gpr/gpr.h"
|
|
#include "mca/ns/ns.h"
|
|
#include "mca/ras/base/ras_base_node.h"
|
|
|
|
static void orte_ras_base_node_construct(orte_ras_node_t* node)
|
|
{
|
|
node->node_name = NULL;
|
|
node->node_arch = NULL;
|
|
node->node_cellid = 0;
|
|
node->node_state = ORTE_NODE_STATE_UNKNOWN;
|
|
node->node_slots = 0;
|
|
node->node_slots_alloc = 0;
|
|
node->node_slots_inuse = 0;
|
|
node->node_slots_max = 0;
|
|
}
|
|
|
|
static void orte_ras_base_node_destruct(orte_ras_node_t* node)
|
|
{
|
|
if (NULL != node->node_name) {
|
|
free(node->node_name);
|
|
}
|
|
if (NULL != node->node_arch) {
|
|
free(node->node_arch);
|
|
}
|
|
}
|
|
|
|
|
|
OBJ_CLASS_INSTANCE(
|
|
orte_ras_node_t,
|
|
opal_list_item_t,
|
|
orte_ras_base_node_construct,
|
|
orte_ras_base_node_destruct);
|
|
|
|
|
|
/*
|
|
* Query the registry for all available nodes
|
|
*/
|
|
|
|
int orte_ras_base_node_query(opal_list_t* nodes)
|
|
{
|
|
size_t i, cnt;
|
|
orte_gpr_value_t** values;
|
|
int rc;
|
|
|
|
/* query all node entries */
|
|
rc = orte_gpr.get(
|
|
ORTE_GPR_KEYS_OR|ORTE_GPR_TOKENS_OR,
|
|
ORTE_NODE_SEGMENT,
|
|
NULL,
|
|
NULL,
|
|
&cnt,
|
|
&values);
|
|
if(ORTE_SUCCESS != rc) {
|
|
ORTE_ERROR_LOG(rc);
|
|
return rc;
|
|
}
|
|
|
|
/* parse the response */
|
|
for(i=0; i<cnt; i++) {
|
|
orte_gpr_value_t* value = values[i];
|
|
orte_ras_node_t* node = OBJ_NEW(orte_ras_node_t);
|
|
size_t k;
|
|
|
|
for(k=0; k<value->cnt; k++) {
|
|
orte_gpr_keyval_t* keyval = value->keyvals[k];
|
|
if(strcmp(keyval->key, ORTE_NODE_NAME_KEY) == 0) {
|
|
node->node_name = strdup(keyval->value.strptr);
|
|
continue;
|
|
}
|
|
if(strcmp(keyval->key, ORTE_NODE_ARCH_KEY) == 0) {
|
|
node->node_arch = strdup(keyval->value.strptr);
|
|
continue;
|
|
}
|
|
if(strcmp(keyval->key, ORTE_NODE_STATE_KEY) == 0) {
|
|
node->node_state = keyval->value.node_state;
|
|
continue;
|
|
}
|
|
if(strcmp(keyval->key, ORTE_NODE_SLOTS_KEY) == 0) {
|
|
node->node_slots = keyval->value.size;
|
|
continue;
|
|
}
|
|
if(strncmp(keyval->key, ORTE_NODE_SLOTS_ALLOC_KEY, strlen(ORTE_NODE_SLOTS_ALLOC_KEY)) == 0) {
|
|
node->node_slots_inuse += keyval->value.size;
|
|
continue;
|
|
}
|
|
if(strcmp(keyval->key, ORTE_NODE_SLOTS_MAX_KEY) == 0) {
|
|
node->node_slots_max = keyval->value.size;
|
|
continue;
|
|
}
|
|
if(strcmp(keyval->key, ORTE_CELLID_KEY) == 0) {
|
|
node->node_cellid = keyval->value.cellid;
|
|
continue;
|
|
}
|
|
}
|
|
opal_list_append(nodes, &node->super);
|
|
}
|
|
return ORTE_SUCCESS;
|
|
}
|
|
|
|
/*
|
|
* Query the registry for all nodes allocated to a specified job
|
|
*/
|
|
int orte_ras_base_node_query_alloc(opal_list_t* nodes, orte_jobid_t jobid)
|
|
{
|
|
char* keys[] = {
|
|
ORTE_NODE_NAME_KEY,
|
|
ORTE_NODE_ARCH_KEY,
|
|
ORTE_NODE_STATE_KEY,
|
|
ORTE_NODE_SLOTS_KEY,
|
|
ORTE_NODE_SLOTS_ALLOC_KEY,
|
|
ORTE_NODE_SLOTS_MAX_KEY,
|
|
ORTE_CELLID_KEY,
|
|
NULL
|
|
};
|
|
size_t i, cnt;
|
|
orte_gpr_value_t** values;
|
|
char* jobid_str;
|
|
int rc;
|
|
|
|
if(ORTE_SUCCESS != (rc = orte_ns.convert_jobid_to_string(&jobid_str, jobid))) {
|
|
ORTE_ERROR_LOG(rc);
|
|
return rc;
|
|
}
|
|
asprintf(&keys[4], "%s-%s", ORTE_NODE_SLOTS_ALLOC_KEY, jobid_str);
|
|
free(jobid_str);
|
|
|
|
/* query selected node entries */
|
|
rc = orte_gpr.get(
|
|
ORTE_GPR_KEYS_OR|ORTE_GPR_TOKENS_OR,
|
|
ORTE_NODE_SEGMENT,
|
|
NULL,
|
|
keys,
|
|
&cnt,
|
|
&values);
|
|
if(ORTE_SUCCESS != rc) {
|
|
ORTE_ERROR_LOG(rc);
|
|
return rc;
|
|
}
|
|
|
|
/* parse the response */
|
|
for(i=0; i<cnt; i++) {
|
|
orte_gpr_value_t* value = values[i];
|
|
orte_ras_node_t* node = OBJ_NEW(orte_ras_node_t);
|
|
size_t k;
|
|
|
|
for(k=0; k<value->cnt; k++) {
|
|
orte_gpr_keyval_t* keyval = value->keyvals[k];
|
|
if(strcmp(keyval->key, ORTE_NODE_NAME_KEY) == 0) {
|
|
node->node_name = strdup(keyval->value.strptr);
|
|
continue;
|
|
}
|
|
if(strcmp(keyval->key, ORTE_NODE_ARCH_KEY) == 0) {
|
|
node->node_arch = strdup(keyval->value.strptr);
|
|
continue;
|
|
}
|
|
if(strcmp(keyval->key, ORTE_NODE_STATE_KEY) == 0) {
|
|
node->node_state = keyval->value.node_state;
|
|
continue;
|
|
}
|
|
if(strcmp(keyval->key, ORTE_NODE_SLOTS_KEY) == 0) {
|
|
node->node_slots = keyval->value.size;
|
|
continue;
|
|
}
|
|
if(strncmp(keyval->key, ORTE_NODE_SLOTS_ALLOC_KEY, strlen(ORTE_NODE_SLOTS_ALLOC_KEY)) == 0) {
|
|
node->node_slots_inuse += keyval->value.size;
|
|
node->node_slots_alloc += keyval->value.size;
|
|
continue;
|
|
}
|
|
if(strcmp(keyval->key, ORTE_NODE_SLOTS_MAX_KEY) == 0) {
|
|
node->node_slots_max = keyval->value.size;
|
|
continue;
|
|
}
|
|
if(strcmp(keyval->key, ORTE_CELLID_KEY) == 0) {
|
|
node->node_cellid = keyval->value.cellid;
|
|
continue;
|
|
}
|
|
}
|
|
/* in case we get back more than we asked for */
|
|
if(node->node_slots_inuse == 0) {
|
|
OBJ_RELEASE(node);
|
|
continue;
|
|
}
|
|
opal_list_append(nodes, &node->super);
|
|
}
|
|
return ORTE_SUCCESS;
|
|
}
|
|
|
|
/*
|
|
* Add the specified node definitions to the registry
|
|
*/
|
|
int orte_ras_base_node_insert(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++) {
|
|
orte_gpr_value_t* value = values[i] = OBJ_NEW(orte_gpr_value_t);
|
|
if (NULL == value) {
|
|
for (j=0; j < i; j++) {
|
|
OBJ_RELEASE(values[j]);
|
|
}
|
|
free(values);
|
|
ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
|
|
return ORTE_ERR_OUT_OF_RESOURCE;
|
|
}
|
|
|
|
value->addr_mode = ORTE_GPR_OVERWRITE | ORTE_GPR_TOKENS_AND;
|
|
value->segment = strdup(ORTE_NODE_SEGMENT);
|
|
value->cnt = 6;
|
|
value->keyvals = (orte_gpr_keyval_t**)malloc(value->cnt*sizeof(orte_gpr_keyval_t*));
|
|
if (NULL == value->keyvals) {
|
|
for (j=0; j < i; j++) {
|
|
OBJ_RELEASE(values[j]);
|
|
}
|
|
free(values);
|
|
ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
|
|
return ORTE_ERR_OUT_OF_RESOURCE;
|
|
}
|
|
|
|
for (j=0; j < value->cnt; j++) {
|
|
value->keyvals[j] = OBJ_NEW(orte_gpr_keyval_t);
|
|
if (NULL == value->keyvals[j]) {
|
|
for (j=0; j <= i; j++) {
|
|
OBJ_RELEASE(values[j]);
|
|
}
|
|
free(values);
|
|
ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
|
|
return ORTE_ERR_OUT_OF_RESOURCE;
|
|
}
|
|
}
|
|
}
|
|
|
|
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)) {
|
|
orte_gpr_value_t* value = values[i];
|
|
node = (orte_ras_node_t*)item;
|
|
|
|
j = 0;
|
|
(value->keyvals[j])->key = strdup(ORTE_NODE_NAME_KEY);
|
|
(value->keyvals[j])->type = ORTE_STRING;
|
|
(value->keyvals[j])->value.strptr = strdup(node->node_name);
|
|
|
|
++j;
|
|
(value->keyvals[j])->key = strdup(ORTE_NODE_ARCH_KEY);
|
|
(value->keyvals[j])->type = ORTE_STRING;
|
|
if (NULL != node->node_arch) {
|
|
(value->keyvals[j])->value.strptr = strdup(node->node_arch);
|
|
} else {
|
|
(value->keyvals[j])->value.strptr = strdup("");
|
|
}
|
|
|
|
++j;
|
|
(value->keyvals[j])->key = strdup(ORTE_NODE_STATE_KEY);
|
|
(value->keyvals[j])->type = ORTE_NODE_STATE;
|
|
(value->keyvals[j])->value.node_state = node->node_state;
|
|
|
|
++j;
|
|
(value->keyvals[j])->key = strdup(ORTE_CELLID_KEY);
|
|
(value->keyvals[j])->type = ORTE_CELLID;
|
|
(value->keyvals[j])->value.cellid = node->node_cellid;
|
|
|
|
++j;
|
|
(value->keyvals[j])->key = strdup(ORTE_NODE_SLOTS_KEY);
|
|
(value->keyvals[j])->type = ORTE_SIZE;
|
|
(value->keyvals[j])->value.size = node->node_slots;
|
|
|
|
++j;
|
|
(value->keyvals[j])->key = strdup(ORTE_NODE_SLOTS_MAX_KEY);
|
|
(value->keyvals[j])->type = ORTE_SIZE;
|
|
(value->keyvals[j])->value.size = node->node_slots_max;
|
|
|
|
/* setup index/keys for this node */
|
|
rc = orte_schema.get_node_tokens(&value->tokens, &value->num_tokens, node->node_cellid, node->node_name);
|
|
if (ORTE_SUCCESS != rc) {
|
|
ORTE_ERROR_LOG(rc);
|
|
for (j=0; j <= i; j++) {
|
|
OBJ_RELEASE(values[j]);
|
|
}
|
|
free(values);
|
|
return rc;
|
|
}
|
|
}
|
|
|
|
/* try the insert */
|
|
if (ORTE_SUCCESS != (rc = orte_gpr.put(num_values, values))) {
|
|
ORTE_ERROR_LOG(rc);
|
|
}
|
|
|
|
for (j=0; j < num_values; j++) {
|
|
OBJ_RELEASE(values[j]);
|
|
}
|
|
free(values);
|
|
return rc;
|
|
}
|
|
|
|
/*
|
|
* Delete the specified nodes from the registry
|
|
*/
|
|
int orte_ras_base_node_delete(opal_list_t* nodes)
|
|
{
|
|
opal_list_item_t* item;
|
|
int rc;
|
|
size_t num_values, num_tokens;
|
|
orte_ras_node_t* node;
|
|
char** tokens;
|
|
|
|
num_values = opal_list_get_size(nodes);
|
|
if (0 >= num_values) {
|
|
ORTE_ERROR_LOG(ORTE_ERR_BAD_PARAM);
|
|
return ORTE_ERR_BAD_PARAM;
|
|
}
|
|
|
|
for(item = opal_list_get_first(nodes);
|
|
item != opal_list_get_end(nodes);
|
|
item = opal_list_get_next(item)) {
|
|
node = (orte_ras_node_t*)item;
|
|
|
|
/* setup index/keys for this node */
|
|
rc = orte_schema.get_node_tokens(&tokens, &num_tokens, node->node_cellid, node->node_name);
|
|
if (ORTE_SUCCESS != rc) {
|
|
ORTE_ERROR_LOG(rc);
|
|
return rc;
|
|
}
|
|
|
|
rc = orte_gpr.delete_entries(
|
|
ORTE_GPR_TOKENS_AND,
|
|
ORTE_NODE_SEGMENT,
|
|
tokens,
|
|
NULL);
|
|
if(ORTE_SUCCESS != rc) {
|
|
ORTE_ERROR_LOG(rc);
|
|
return rc;
|
|
}
|
|
}
|
|
return ORTE_SUCCESS;
|
|
}
|
|
|
|
/*
|
|
* Assign the allocated slots on the specified nodes to the
|
|
* indicated jobid.
|
|
*/
|
|
int orte_ras_base_node_assign(opal_list_t* nodes, orte_jobid_t jobid)
|
|
{
|
|
opal_list_item_t* item;
|
|
orte_gpr_value_t **values;
|
|
int rc;
|
|
size_t num_values, i, j;
|
|
orte_ras_node_t* node;
|
|
char* jobid_str;
|
|
|
|
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++) {
|
|
values[i] = OBJ_NEW(orte_gpr_value_t);
|
|
if (NULL == values[i]) {
|
|
for (j=0; j < i; j++) {
|
|
OBJ_RELEASE(values[j]);
|
|
}
|
|
free(values);
|
|
ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
|
|
return ORTE_ERR_OUT_OF_RESOURCE;
|
|
}
|
|
|
|
values[i]->addr_mode = ORTE_GPR_OVERWRITE | ORTE_GPR_TOKENS_AND;
|
|
values[i]->segment = strdup(ORTE_NODE_SEGMENT);
|
|
values[i]->cnt = 1;
|
|
values[i]->keyvals = (orte_gpr_keyval_t**)malloc(sizeof(orte_gpr_keyval_t*));
|
|
if (NULL == values[i]->keyvals) {
|
|
for (j=0; j < i; j++) {
|
|
OBJ_RELEASE(values[j]);
|
|
}
|
|
free(values);
|
|
ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
|
|
return ORTE_ERR_OUT_OF_RESOURCE;
|
|
}
|
|
|
|
values[i]->keyvals[0] = OBJ_NEW(orte_gpr_keyval_t);
|
|
if (NULL == values[i]->keyvals[0]) {
|
|
for (j=0; j < i; j++) {
|
|
OBJ_RELEASE(values[j]);
|
|
}
|
|
free(values);
|
|
ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
|
|
return ORTE_ERR_OUT_OF_RESOURCE;
|
|
}
|
|
}
|
|
|
|
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)) {
|
|
int rc;
|
|
node = (orte_ras_node_t*)item;
|
|
|
|
if(node->node_slots_alloc == 0)
|
|
continue;
|
|
if(ORTE_SUCCESS != (rc = orte_ns.convert_jobid_to_string(&jobid_str, jobid))) {
|
|
ORTE_ERROR_LOG(rc);
|
|
return rc;
|
|
}
|
|
|
|
/* 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);
|
|
for (j=0; j < num_values; j++) {
|
|
OBJ_RELEASE(values[j]);
|
|
}
|
|
free(values);
|
|
return rc;
|
|
}
|
|
|
|
/* setup node key/value pairs */
|
|
asprintf(&((values[i]->keyvals[0])->key), "%s-%s", ORTE_NODE_SLOTS_ALLOC_KEY, jobid_str);
|
|
free(jobid_str);
|
|
|
|
(values[i]->keyvals[0])->type = ORTE_SIZE;
|
|
(values[i]->keyvals[0])->value.size = node->node_slots_alloc;
|
|
}
|
|
|
|
/* try the insert */
|
|
if (ORTE_SUCCESS != (rc = orte_gpr.put(num_values, values))) {
|
|
ORTE_ERROR_LOG(rc);
|
|
}
|
|
|
|
for (j=0; j < num_values; j++) {
|
|
OBJ_RELEASE(values[j]);
|
|
}
|
|
free(values);
|
|
|
|
return rc;
|
|
}
|
|
|