23 ianuarie 2008 20:37OK, so I got client utilities set up on a WinXP Pro (32-bit) PC and it is connected to the cluster and I can see the jobs in the job scheduler and tasks within eacg job (had to be an admin on the cluster for that) but when I submit a job from this PC the job immediately fails and all the tasks within the job show up as status of "Queued". I can submit a an identical job (with the same job template) from the head node and it runs fine.
Any suggests here would be greatly appreciated. Trying to demo this functionality soon. Thanks!
- Mutat de Josh BarnardModerator 25 martie 2009 20:31 (Moved from Windows HPC Server Deployment, Management, and Administration to Windows HPC Server Job Submission and Scheduling)
23 ianuarie 2008 22:44Just guessing: did you check that the submitter's credentials are registered as a cluster user in the admin console? Is there any error message? Can you try a task view / job view from the command line? Can you re-direct stdout and stderr to look for error messages?Did you have a look at the log files?G.
24 ianuarie 2008 14:57Thanks for the sanity check. The submitter is registered as a cluster user and administrator, definitely. There is no error message, hence my confusion. The job lists are failed but the individual tasks are techinically failed but simply queued. There are stdout and stderr files associated with each task but they do not get written since the job fails immediately. I do not use the command line much, haven't tried to much and doesn't seem useful for the casual users I am trying to pull in to using it.
What log files are you refering to? I would be interested in these but I don't think I am aware of them.
25 martie 2009 20:31Moderator
Just saw this was never answered . . . if you're stil experiencing the problem, try posting again in the Scheduler forums.
- Marcat ca răspuns de Josh BarnardModerator 27 martie 2009 18:03