int main(int argc , char ** argv) {
const int queue_timeout = 180;
const int submit_timeout = 180;
const int status_timeout = 180;
const int number_of_jobs = 250;
const int submit_threads = number_of_jobs / 10 ;
const int status_threads = number_of_jobs + 1;
const char * job = util_alloc_abs_path(argv[1]);
rng_type * rng = rng_alloc( MZRAN , INIT_CLOCK );
test_work_area_type * work_area = test_work_area_alloc("job_queue");
job_type **jobs = alloc_jobs( rng , number_of_jobs , job);
job_queue_type * queue = job_queue_alloc(number_of_jobs, "OK", "ERROR");
queue_driver_type * driver = queue_driver_alloc_local();
job_queue_manager_type * queue_manager = job_queue_manager_alloc( queue );
job_queue_set_driver(queue, driver);
job_queue_manager_start_queue(queue_manager, 0, false , true);
{
thread_pool_type * status_pool = thread_pool_alloc( status_threads , true );
thread_pool_type * submit_pool = thread_pool_alloc( submit_threads , true );
submit_jobs( queue , number_of_jobs , jobs , submit_pool );
status_jobs( queue , number_of_jobs , jobs , status_pool );
if (!thread_pool_try_join( submit_pool , submit_timeout ))
util_exit("Joining submit pool failed \n");
thread_pool_free( submit_pool );
job_queue_submit_complete(queue);
if (!thread_pool_try_join( status_pool , status_timeout))
util_exit("Joining status pool failed \n");
thread_pool_free( status_pool );
}
if (!job_queue_manager_try_wait(queue_manager , queue_timeout))
util_exit("job_queue never completed \n");
job_queue_manager_free(queue_manager);
job_queue_free(queue);
queue_driver_free(driver);
check_jobs( number_of_jobs , jobs );
test_work_area_free(work_area);
rng_free( rng );
}
int main(int argc, char *argv[])
{
char diagnosis[DRMAA_ERROR_STRING_BUFFER];
char *s, jobid[100];
int drmaa_errno, i;
int ret = 0;
struct timeval start_s, finish_s, wait_s;
if (argc<2) {
usage();
return 1;
}
i = 1;
do {
if (!strcmp("-help", argv[i]) ||
!strcmp("-h", argv[i])) {
usage();
return 0;
} else if (!strcmp("-jobs", argv[i])) {
i++;
if (argc < i+1) {
usage();
return 1;
}
njobs = atoi(argv[i]);
i++;
} else if (!strcmp("-native", argv[i])) {
i++;
if (argc < i+1) {
usage();
return 1;
}
native_spec = argv[i];
i++;
} else if (!strcmp("-threads", argv[i])) {
if (argc < i+1) {
usage();
return 1;
}
i++;
nthreads = atoi(argv[i]);
i++;
} else if (!strcmp("-quiet", argv[i])) {
i++;
if (argc < i+1) {
usage();
return 1;
}
if (!strcmp("yes", argv[i]) || !strcmp("y", argv[i]))
quiet = 1;
else if (!strcmp("no", argv[i]) || !strcmp("n", argv[i]))
quiet = 0;
else {
usage();
return 1;
}
i++;
} else if (!strcmp("-wait", argv[i])) {
i++;
if (argc < i+1) {
usage();
return 1;
}
if (!strcmp("yes", argv[i]) || !strcmp("y", argv[i]))
dowait = 1;
else if (!strcmp("no", argv[i]) || !strcmp("n", argv[i]))
dowait = 0;
else {
usage();
return 1;
}
i++;
} else if (!strcmp("-scenario", argv[i])) {
i++;
if (argc < i+1) {
usage();
return 1;
}
s = strchr(argv[i], '.');
*s = '\0';
if (strcmp("queue", argv[i]) && strcmp("type", argv[i]) &&
strcmp("number", argv[i]) && strcmp("pe", argv[i])) {
usage();
return 1;
}
scenario = strdup(argv[i]);
s++;
if (strcmp("hostgroup", s) && strcmp("resource", s) &&
strcmp("none", s) && strcmp("softresource", s) &&
strcmp("softhostgroup", s)) {
usage();
return 1;
}
site_b = strdup(s);
i++;
} else {
job_path = argv[i];
i++;
if (job_path[0]=='-') {
usage();
return 1;
}
if (argv[i]) {
job_args = &argv[i];
}
}
} while (i < argc && !job_path);
if (!job_path) {
usage();
return 1;
}
#if 0
printf("job_path: \"%s\"\n", job_path);
printf("njobs: %d\n", njobs);
printf("nthreads: %d\n", nthreads);
printf("native: %s\n", native_spec);
printf("dowait: %s\n", dowait?"yes":"no");
printf("quiet: %s\n", quiet?"yes":"no");
printf("scenario: %s\n", scenario?scenario:"<no such>");
printf("site_b: %s\n", site_b?site_b:"<no such>");
printf("1st arg: %s\n", job_args?job_args[0]:"<noargs>");
#endif
if (drmaa_init(NULL, diagnosis, sizeof(diagnosis)-1) != DRMAA_ERRNO_SUCCESS) {
fprintf(stderr, "drmaa_init() failed: %s\n", diagnosis);
return 1;
}
get_gmt(&start_s);
if (!scenario) {
if (!getuid()) {
fprintf(stderr, "switching to ah114088:gridware\n");
setegid(339);
seteuid(115088);
}
if (!(jt = create_job_template(job_path, NULL, 0))) {
fprintf(stderr, "create_sleeper_job_template() failed\n");
return 1;
}
if (nthreads==1) {
if (submit_jobs(&argv[i]))
return 1;
} else {
pthread_t *ids = NULL;
ids = (pthread_t *)malloc(sizeof (pthread_t) * nthreads);
for (i = 0; i < nthreads; i++) {
if (pthread_create(&ids[i], NULL, submit_jobs, NULL)) {
fprintf(stderr, "pthread_create() failed: %s\n", strerror(errno));
free(ids);
return 1;
}
}
for (i = 0; i < nthreads; i++) {
pthread_join(ids[i], NULL);
}
}
drmaa_delete_job_template(jt, NULL, 0);
if (!getuid()) {
fprintf(stderr, "switching to root:root\n");
seteuid(0);
setegid(0);
}
} else {
if (submit_by_project("project1") || submit_by_project("project2") ||
submit_by_project("project3") || submit_by_project("project4"))
return 1;
}
get_gmt(&finish_s);
printf("submission took %8.3f seconds\n", DELTA_SECONDS(start_s, finish_s));
if (dowait) {
int success = 1;
for (i=0; i<njobs * nthreads; i++) {
int stat;
int aborted, exited, exit_status, signaled;
drmaa_errno = drmaa_wait(DRMAA_JOB_IDS_SESSION_ANY, jobid, sizeof(jobid)-1,
&stat, DRMAA_TIMEOUT_WAIT_FOREVER, NULL, diagnosis, sizeof(diagnosis)-1);
if (drmaa_errno != DRMAA_ERRNO_SUCCESS) {
fprintf(stderr, "drmaa_wait() failed: %s\n", diagnosis);
return 1;
}
/*
* report how job finished
*/
drmaa_wifaborted(&aborted, stat, NULL, 0);
if (aborted) {
printf("job \"%s\" never ran\n", jobid);
success = 0;
} else {
drmaa_wifexited(&exited, stat, NULL, 0);
if (exited) {
drmaa_wexitstatus(&exit_status, stat, NULL, 0);
if (exit_status != 0) {
success = 0;
printf("job \"%s\" with exit status %d\n", jobid, exit_status);
} else {
if (!quiet)
printf("job \"%s\" finished regularly\n", jobid);
}
} else {
success = 0;
drmaa_wifsignaled(&signaled, stat, NULL, 0);
if (signaled) {
char termsig[DRMAA_SIGNAL_BUFFER+1];
drmaa_wtermsig(termsig, DRMAA_SIGNAL_BUFFER, stat, NULL, 0);
printf("job \"%s\" finished due to signal %s\n", jobid, termsig);
} else
printf("job \"%s\" finished with unclear conditions\n", jobid);
}
}
}
if (!success)
ret = 1;
get_gmt(&wait_s);
printf("wait took %8.3f seconds\n", DELTA_SECONDS(finish_s, wait_s));
printf("jobs took %8.3f seconds\n", DELTA_SECONDS(start_s, wait_s));
}
if (drmaa_exit(diagnosis, sizeof(diagnosis)-1) != DRMAA_ERRNO_SUCCESS) {
fprintf(stderr, "drmaa_exit() failed: %s\n", diagnosis);
return 1;
}
return ret;
}
