RTI Server invokes the same job in both server

[williamswe]

Hi All,

We have 2 DataStage servers with identical jobs/configuration for High Availability. The RTI server should read parameters & time-to-run from the schedule table and run the job on one of servers ( Job Status will be updated once the job is done) .
Today, our RTI Server( 7.5.0.1 ) invokes a job in one server and after 1 min it invokes the same job in another server. The job normally run 10-30 secs.

Do you guys have any idea on how could this happen..

Thanks a lot,
William

[roy]

Hi and welcome aboard ,

So if understand you correcly your job got run on both servers instead of being run only once on one of them?

If so contact your support (please post a reply and rport your findings/solution)

Was there anything else that went not as expected?

Hi All,

We have 2 DataStage servers with identical jobs/configuration for High Availability. The RTI server should read parameters & time-to-run from the schedule table and run the job on one of servers ( Job Status will be updated once the job is done) .
Today, our RTI Server( 7.5.0.1 ) invokes a job in one server and after 1 min it invokes the same job in another server. The job normally run 10-30 secs.

Do you guys have any idea on how could this happen..

Thanks a lot,
William

[williamswe]

Thx Roy,

We have already contacted the support team and the case is being investigated now.
The initial findings are:
1) we are quite sure that only one of the RTI/TX servers was running(active) at that time.
2) RTI Localhost_Access_log shows "there was only one service call to that particular job". some how it was kicked off again on another server ( 1 min after the first one was completed )
3) director logs show both jobs have invocation # attached to the job name.

Thx
william

[williamswe]

Case Closed!!

It turned out that we didn't pay attention to details.
RTI/Tx was working fine...data duplication was caused by some mistakes in the job design.

RTI local acess log shows time in 24 hr format :
one service call was on
28/Mar/2006:00:54:26 +0800
another at
28/Mar/2006:12:52:49 +0800

but the DS director log shows time in am/pm format:
on server1, job run at
12:52:38 PM
on server2, at
12:53:59 AM

i think the AM/PM is too easy to be ignored sometimes, is there anyway to change it 24 hr format???

[ray.wurlod]

It should be picked up from your Windows regional settings.