Hi,
I have a problem in scheduling jobs in DataStage. The DataStage
server version 3.5 runs on HP-UX machine.
I have created a user id by the name of dstage in the HP-UX. I
developed all my DataStage jobs using the dstage id in a particular
project. I managed to compile and execute the jobs. Unfortunately, when
I tried to schedule a job in the DataStage Director, it did not run as
scheduled.
Before I schedule the job in the DataStage Director, I have added the
dstage user id in the cron.allow and at.allow file. In order to
ensure that the dstage user can schedule a cronjob, I have created a
simple Unix script and scheduled the script execution in the Unix
crontab. The Unix script managed to execute as scheduled.
The funny thing was that the job execution worked when I tried to
schedule the job using root id in the DataStage Director.
I have checked on the Unix file permissions, and it seems the job
permissions are fine and should be able to access by dstage user.
Anyway, the dstage user can run the job from the Director.
Could anyone figure out my problem? I would really appreciate your
advice here.
Regards,
Melvin
Could not schedule jobs
Hi Melvin,
try to give read and execute right on the /usr/spool/cron/atjobs for datastage user
Regards
Riccardo Tani
Staff Consultant
Unita Servizi alle Imprese
Datamat, Enabling Your Future
Rome, Italy
e-mail: rtani@datamat.it
----- Original Message -----
From: "Melvin Thong"
To:
Sent: Wednesday, May 16, 2001 3:56 PM
Subject: Could not schedule jobs
> Hi,
>
> I have a problem in scheduling jobs in DataStage. The DataStage
> server version 3.5 runs on HP-UX machine.
>
> I have created a user id by the name of dstage in the HP-UX. I
> developed all my DataStage jobs using the dstage id in a particular
> project. I managed to compile and execute the jobs. Unfortunately,
> when I tried to schedule a job in the DataStage Director, it did not
> run as scheduled.
>
> Before I schedule the job in the DataStage Director, I have added
> the dstage user id in the cron.allow and at.allow file. In order
> to ensure that the dstage user can schedule a cronjob, I have
> created a simple Unix script and scheduled the script execution in the
> Unix crontab. The Unix script managed to execute as scheduled.
>
> The funny thing was that the job execution worked when I tried to
> schedule the job using root id in the DataStage Director.
>
> I have checked on the Unix file permissions, and it seems the job
> permissions are fine and should be able to access by dstage user.
> Anyway, the dstage user can run the job from the Director.
>
> Could anyone figure out my problem? I would really appreciate your
> advice here.
>
> Regards,
> Melvin
>
>
try to give read and execute right on the /usr/spool/cron/atjobs for datastage user
Regards
Riccardo Tani
Staff Consultant
Unita Servizi alle Imprese
Datamat, Enabling Your Future
Rome, Italy
e-mail: rtani@datamat.it
----- Original Message -----
From: "Melvin Thong"
To:
Sent: Wednesday, May 16, 2001 3:56 PM
Subject: Could not schedule jobs
> Hi,
>
> I have a problem in scheduling jobs in DataStage. The DataStage
> server version 3.5 runs on HP-UX machine.
>
> I have created a user id by the name of dstage in the HP-UX. I
> developed all my DataStage jobs using the dstage id in a particular
> project. I managed to compile and execute the jobs. Unfortunately,
> when I tried to schedule a job in the DataStage Director, it did not
> run as scheduled.
>
> Before I schedule the job in the DataStage Director, I have added
> the dstage user id in the cron.allow and at.allow file. In order
> to ensure that the dstage user can schedule a cronjob, I have
> created a simple Unix script and scheduled the script execution in the
> Unix crontab. The Unix script managed to execute as scheduled.
>
> The funny thing was that the job execution worked when I tried to
> schedule the job using root id in the DataStage Director.
>
> I have checked on the Unix file permissions, and it seems the job
> permissions are fine and should be able to access by dstage user.
> Anyway, the dstage user can run the job from the Director.
>
> Could anyone figure out my problem? I would really appreciate your
> advice here.
>
> Regards,
> Melvin
>
>
I just encountered the same issue on one of my systems,
it works as "run now" on director
it works as a "root" user scheduled
it doesnt work as a normal user scheduled
The permissions issue was apparent, but easily found
In working with DataStage support, I edited the dsr_sched.sh (spelling?) in whatever directory you installed the project into and made a simple change to add a system trace.
After this edit, scheduling started to work. I removed the change and the scheduled job still ran to completion. Editting that file seemed to make the difference. Why?, I have no idea. Hope it helps.
it works as "run now" on director
it works as a "root" user scheduled
it doesnt work as a normal user scheduled
The permissions issue was apparent, but easily found
In working with DataStage support, I edited the dsr_sched.sh (spelling?) in whatever directory you installed the project into and made a simple change to add a system trace.
After this edit, scheduling started to work. I removed the change and the scheduled job still ran to completion. Editting that file seemed to make the difference. Why?, I have no idea. Hope it helps.
Hi Bob,
Thanks for your suggestion on the scheduling problem.
By the way, I would like to clarify with you what does it mean by adding
a system trace in the dsr_sched.sh? Is it you just put in a simple Unix
echo $... command into the dsr_sched.sh?
Thanks again.
Regards,
Melvin
----- Original Message -----
From:
Date: Thursday, May 17, 2001 5:55 am
Subject: Re: Could not schedule jobs
>
> I just encountered the same issue on one of my systems,
> it works as "run now" on director
> it works as a "root" user scheduled
> it doesnt work as a normal user scheduled
> The permissions issue was apparent, but easily found
>
> In working with DataStage support, I edited the dsr_sched.sh
> (spelling?)in whatever directory you installed the project into
> and made a simple
> change to add a system trace.
>
> After this edit, scheduling started to work. I removed the change
> and the
> scheduled job still ran to completion. Editting that file seemed
> to make
> the difference. Why?, I have no idea. Hope it helps.
>
>
Thanks for your suggestion on the scheduling problem.
By the way, I would like to clarify with you what does it mean by adding
a system trace in the dsr_sched.sh? Is it you just put in a simple Unix
echo $... command into the dsr_sched.sh?
Thanks again.
Regards,
Melvin
----- Original Message -----
From:
Date: Thursday, May 17, 2001 5:55 am
Subject: Re: Could not schedule jobs
>
> I just encountered the same issue on one of my systems,
> it works as "run now" on director
> it works as a "root" user scheduled
> it doesnt work as a normal user scheduled
> The permissions issue was apparent, but easily found
>
> In working with DataStage support, I edited the dsr_sched.sh
> (spelling?)in whatever directory you installed the project into
> and made a simple
> change to add a system trace.
>
> After this edit, scheduling started to work. I removed the change
> and the
> scheduled job still ran to completion. Editting that file seemed
> to make
> the difference. Why?, I have no idea. Hope it helps.
>
>
Hi, Melvin
The trace Datastage support asked me to put into the dsr_sched.sh script was a "truss" statement around the command piped to an output file. (eg. truss $uvhome...... 2>/tmp/trussout). I assume you have verified that the Datastage job is getting to the "at queue. We routinely grabbed the "at" script that was sent to the queue and attempted to execute it directly on the UNIX system and received an error unable to open a file in the uvhome/libs directory. The explanation from here gets too voluminous to continue, but Ascential Support should have a good dialogue on the problem. Bob
The trace Datastage support asked me to put into the dsr_sched.sh script was a "truss" statement around the command piped to an output file. (eg. truss $uvhome...... 2>/tmp/trussout). I assume you have verified that the Datastage job is getting to the "at queue. We routinely grabbed the "at" script that was sent to the queue and attempted to execute it directly on the UNIX system and received an error unable to open a file in the uvhome/libs directory. The explanation from here gets too voluminous to continue, but Ascential Support should have a good dialogue on the problem. Bob