Multiple Job Servers and Max No of Processes

dataservices administration
1

I’ve inherited a data services environment and trying to get things up to speed. I’ve usually only had a single instance of the job server running on each physical server, but here they have a separate instance for each repository. Even in production they have 3 separate repositories and thus 3 job servers. From what I understand and have read here there is no real value in this but I guess not much harm either. But I thought I’d run it by you all.

Another issue is the max_no_of_processes setting. I was always told it was per job server, and hence a reason why you should only have one job server per server. But reading here I see it’s per job, so if you have a lot of jobs that may not always overlap with the same timing it might be hard to optimize. Any advice?

Thanks.

craiggsmith (BOB member since 2009-06-01)

2

I don’t think there is any value in creating multiple job servers within the same job service.

With regard to the number of engines it is indeed job specific. So a job that tries to run 20 Dataflows in parallel will only get 7 if the max engines is set to 8 (the default). If 20 jobs run at the same time and each job has 20 Dataflows then you’ll see 160 al_engine processes (if the O/S doesn’t just up and die). The job itself counts as one engine.

eganjp :us: (BOB member since 2007-09-12)