1
1

Return an informative error message if the user requests a migration of a job that is not capable of it.

C/R Functionality cleanup

This commit was SVN r24307.
Этот коммит содержится в:
Josh Hursey 2011-01-26 15:36:34 +00:00
родитель 8f45fcb429
Коммит 81fd41f811
3 изменённых файлов: 44 добавлений и 0 удалений

Просмотреть файл

@ -124,6 +124,15 @@ int orte_errmgr_hnp_global_module_init(void)
goto cleanup;
}
}
else {
/* Still need the tool listener so we can tell it that we cannot do
* anything if they ask.
*/
if( ORTE_SUCCESS != (ret = orte_errmgr_base_tool_init()) ) {
ORTE_ERROR_LOG(ret);
return ret;
}
}
if( mca_errmgr_hnp_component.autor_enabled ) {
if( ORTE_SUCCESS != (ret = orte_errmgr_hnp_autor_global_module_init()) ) {
@ -153,6 +162,15 @@ int orte_errmgr_hnp_global_module_finalize(void)
goto cleanup;
}
}
else {
/* Still need the tool listener so we can tell it that we cannot do
* anything if they ask.
*/
if( ORTE_SUCCESS != (ret = orte_errmgr_base_tool_finalize()) ) {
ORTE_ERROR_LOG(ret);
return ret;
}
}
if( mca_errmgr_hnp_component.autor_enabled ) {
if( ORTE_SUCCESS != (ret = orte_errmgr_hnp_autor_global_module_finalize()) ) {
@ -255,8 +273,16 @@ int orte_errmgr_hnp_global_predicted_fault(opal_list_t *proc_list,
goto cleanup;
}
}
/*
* If Process migration is not enabled, then return an error the tool
* which will print an appropriate message for the user.
*/
else {
OPAL_OUTPUT_VERBOSE((10, mca_errmgr_hnp_component.super.output_handle,
"errmgr:hnp:predicted_fault() Command line asked for a migration, but it is not enabled\n"));
orte_errmgr_base_migrate_update(ORTE_ERRMGR_MIGRATE_STATE_ERROR);
exit_status = ORTE_ERR_NOT_IMPLEMENTED;
goto cleanup;
}
cleanup:

Просмотреть файл

@ -49,3 +49,10 @@ If you require this feature contact the Open MPI development group.
Error: The Job identified by PID (%d) is currently migrating other processes.
Only one migration request can be processed at a time. Please try again
later.
[err-other]
Error: The Job identified by PID (%d) was not able to migrate processes in this
job. This could be caused by any of the following:
- Invalid node or rank specified
- No processes on the indicated node can by migrated
- Process migration was not enabled for this job. Make sure to indicate
the proper AMCA file: "-am ft-enable-cr-recovery".

Просмотреть файл

@ -600,6 +600,17 @@ static void process_ckpt_update_cmd(orte_process_name_t* sender,
goto cleanup;
}
/*
* If there was an error, display a message and exit
*/
if( ORTE_ERRMGR_MIGRATE_STATE_ERROR == orte_migrate_ckpt_status ) {
opal_show_help("help-orte-migrate.txt", "err-other",
true,
orte_migrate_globals.pid);
exit_status = ORTE_ERROR;
goto cleanup;
}
/*
* If we are to display the status progression
*/