void issue_track(
job *pjob)
{
struct batch_request *preq;
char *pc;
preq = alloc_br(PBS_BATCH_TrackJob);
if (preq == (struct batch_request *)0)
return;
preq->rq_ind.rq_track.rq_hopcount = pjob->ji_wattr[JOB_ATR_hopcount].at_val.at_long;
strcpy(preq->rq_ind.rq_track.rq_jid, pjob->ji_qs.ji_jobid);
strcpy(preq->rq_ind.rq_track.rq_location, server_name);
preq->rq_ind.rq_track.rq_state[0] = pjob->ji_wattr[JOB_ATR_state].at_val.at_char;
pc = pjob->ji_qs.ji_jobid;
while (*pc != '.')
pc++;
issue_to_svr(++pc, preq, NULL);
free_br(preq);
}
static void reissue_to_svr(
struct work_task *pwt)
{
time_t time_now = time(NULL);
char *br_id = pwt->wt_parm1;
batch_request *preq = get_remove_batch_request(br_id);
/* if not timed-out, retry send to remote server */
if (preq != NULL)
{
if (((time_now - preq->rq_time) > PBS_NET_RETRY_LIMIT) ||
(issue_to_svr(preq->rq_host, preq, pwt->wt_parmfunc) != PBSE_NONE))
{
/* either timed-out or got hard error, tell post-function */
pwt->wt_aux = -1; /* seen as error by post function */
pwt->wt_event = -1; /* seen as connection by post func */
if (pwt->wt_parmfunc != NULL)
((void (*)())pwt->wt_parmfunc)(pwt);
}
}
free(pwt->wt_mutex);
free(pwt);
} /* END reissue_to_svr() */
void issue_track(
job *pjob)
{
struct batch_request *preq;
char *pc;
char *sname;
char log_buf[LOCAL_LOG_BUF_SIZE];
if ((pc = strchr(pjob->ji_qs.ji_jobid, '.')) == NULL)
{
snprintf(log_buf, sizeof(log_buf),
"Remote job routing is not compatible with display_job_server_suffix set to false. Cannot track %s",
pjob->ji_qs.ji_jobid);
log_err(-1, __func__, log_buf);
return;
}
sname = pc + 1;
/* do not issue track requests to ourselves */
if (!strcmp(sname, server_name))
{
snprintf(log_buf, sizeof(log_buf),
"%s erroneously called for local job %s",
__func__, pjob->ji_qs.ji_jobid);
log_event(PBSEVENT_JOB, PBS_EVENTCLASS_JOB, __func__, log_buf);
return;
}
preq = alloc_br(PBS_BATCH_TrackJob);
if (preq == NULL)
return;
preq->rq_ind.rq_track.rq_hopcount = pjob->ji_wattr[JOB_ATR_hopcount].at_val.at_long;
strcpy(preq->rq_ind.rq_track.rq_jid, pjob->ji_qs.ji_jobid);
strcpy(preq->rq_ind.rq_track.rq_location, server_name);
preq->rq_ind.rq_track.rq_state[0] = pjob->ji_wattr[JOB_ATR_state].at_val.at_char;
pc = pjob->ji_qs.ji_jobid;
while (*pc != '.')
pc++;
issue_to_svr(++pc, preq, NULL);
free_br(preq);
}
void reissue_to_svr(
struct work_task *pwt)
{
time_t time_now = time(NULL);
char *br_id;
batch_request *preq;
char *serverName = NULL;
if (pwt == NULL)
return;
br_id = (char *)pwt->wt_parm1;
preq = get_remove_batch_request(br_id);
/* if not timed-out, retry send to remote server */
if (preq != NULL)
{
if (preq->rq_host[0] != '\0')
serverName = strdup(preq->rq_host);
else
{
free(pwt->wt_mutex);
free(pwt);
return;
}
if (((time_now - preq->rq_time) > PBS_NET_RETRY_LIMIT) ||
(issue_to_svr(serverName, &preq, pwt->wt_parmfunc) != PBSE_NONE))
{
/* either timed-out or got hard error, tell post-function */
pwt->wt_aux = -1; /* seen as error by post function */
pwt->wt_event = -1; /* seen as connection by post func */
if (pwt->wt_parmfunc != NULL)
(* pwt->wt_parmfunc)(pwt);
}
}
if (serverName)
free(serverName);
free(pwt->wt_mutex);
free(pwt);
} /* END reissue_to_svr() */
/**
* @brief
* reissue_to_svr - recall issue_to_svr() after a delay to retry sending
* a request that failed for a temporary reason
*
* @see
* issue_to_svr
*
* @param[in] pwt - pointer to work structure
*
* @return void
*/
static void
reissue_to_svr(struct work_task *pwt)
{
struct batch_request *preq = pwt->wt_parm1;
int issue_to_svr(char *, struct batch_request *,
void(*)(struct work_task *));
/* if not timed-out, retry send to remote server */
if (((time_now - preq->rq_time) > PBS_NET_RETRY_LIMIT) ||
(issue_to_svr(preq->rq_host, preq, (void(*)(struct work_task *))pwt->wt_parm2) == -1)) {
/* either timed-out or got hard error, tell post-function */
pwt->wt_aux = -1; /* seen as error by post function */
pwt->wt_event = -1; /* seen as connection by post func */
((void (*)())pwt->wt_parm2)(pwt);
}
return;
}
