[torqueusers] Question on req_rdytocommit() and locate_new_job()

Prakash Velayutham velayups at email.uc.edu
Sun Sep 25 16:47:41 MDT 2005


pbs_submit call receives jobid from the server in the connected socket. This is read in the client side using PBSD_rdrpy() call. The next step in the batch request call is req_rdytocommit(). This call sends the same batch request structure as the argument.

req_rdytocommit(struct batch_request *preq)

But inside of the req_rdytocommit() code, I see that the current job is located using locate_new_job() call. This call requires the socket and the jobid. I can see how preq->rq_conn gives the socket number, but I fail to understand how preq->rq_ind.rq_rdytocommit has the correct jobid. When is this set with the server assigned jobid? I can see that preq->rq_reply.brp_un.brp_jid is assigned the server assigned id inside the reply_jobid() routine.

Could someone please explain?


More information about the torqueusers mailing list