cancel
Showing results for 
Search instead for 
Did you mean: 

Large Log files produced by workflow

Former Member
0 Kudos

Are there large log files that can grow quickly when workflow is activated and taken into use? I have set up a workflow for a customer and it has been taken into production for two weeks; this weekend I received a phone call to ask if I knew why they were running out of disk space on their productive system (I don't, I'm no Basis expert). However, it got me thinking if it was anything to do with the workflow activation.

So are there any files I can check to see if they have grown too large too quickly, and if there are, how can I safely trim them down?

Kind Regards,

Tony.

Accepted Solutions (0)

Answers (1)

Answers (1)

0 Kudos

Hi Anthony,

there are a number of SWI* transactions that allow you to look at the various logs produced by Workflow. You may want to see about scheduling the program RSWWWIDE_TOPLEVEL to "cleanup" completed workflows and their logs if the logs are not required for auditing purposes after a period of time.

You could check the Event Trace (txn: SWELS) to ensure it has not been turned on and left on in Prod. (the trace is only small but if it has been left on some time it could be big.) Use txn: SWEL and you wil be able to display/delete it as required.

Workflow uses a number of Transactional RFC calls to manage events. Possibly your basis guys are hording the RFC logs.

Could any of your workflows have gone into a loop? Each pass round the loop will write more data to the log. Too many times around the loop will cause the workflow to fail as the "internal counter" of the log can only handle about 1000 entries.

I think the size impact of each of these things is fairly minor so they should not have caused the space problem in the 2 week timeframe you speak of, but they are things to consider in the long term.

Hope this helps.

Mark

pokrakam
Active Contributor
0 Kudos

C'mon, is this a serious question? Not yours, the one from Basis! THEY should be able to tell you what's filling up their disk, otherwise you might as well shut shop.

OK, to make it a useful response - you should only be concerned if you are triggering a lot of flows. You negelect to mention this rather relevant piece of info. A VERY rough guideline might be over 10,000 WI's/day where you might need to be concerned, look at SWI2_FREQ.

On top of Mark's very good suggestions, I'd also add archiving (tx SARA). But I'd offer a minor correction on the loop theory: This is defined/restricted by SWPA, 10,000 by default. WF cleverly rolls back the loop once it hits that limit so it will only log one pass (at least it does on simple loops).

Oh, also look at report RSWIWADO (outdated because 'how long is a piece of string' means the limits are different for everyone, but still gives you a rough idea).

Cheers,

Mike

Former Member
0 Kudos

Mark and Mike,

Thanks for the answers - very useful!

Don't want to comment about the Basis support of the company in question, but the support is a bit 'thin'...

Anyway, this should help me.

Cheers,

Tony.