[Iplant-api-dev] Jobs intermittently failing at archiving step

[Barthelson, Roger A - (rogerab)]

Jobs do seem to fail sometimes during archiving. Sometimes they will take too long to archive and time out during transfer. I believe Internet transfer problems can contribute to that. I'm not an expert at all on this side of things, but I think we can probably limit this sort of problem sometimes by how we design our Apps. One thing is to remove any large intermediate files that will not be needed by the user. Sometimes you do need these files, and what user's may want is hard to guess sometimes -- in which case you can make it an option. But otherwise, I don't have any great ideas that can help this. There can be anomalies that slow or break transfer, and sometimes it seems like an application is running normally, but you don 't get a final output. In recent experience I've seen some applications just not be able to run successfully through all stages because of some limitation in the data set. I have also seen assemblers go through a lot  of processing, but have no output -- careful inspection showed they just ran out of memory. So what you describe can have multiple causes, including possible transfer problems.

Roger


Sent from my iPad

On Aug 8, 2013, at 1:33 PM, "Jennewein, Douglas M" <Doug.Jennewein at usd.edu<mailto:Doug.Jennewein at usd.edu>> wrote:

We’ve had several jobs run normally, reach the Archiving step, and then occasionally fail, according to the Test Application.

However, the Failed job’s output is visible, for example: /apps-v1/job/19402/output/list, and indicates a successful run.  There are no obvious error messages.

Could we be doing something on our end that (sometimes) causes job archival to fail for otherwise successful runs?

Doug Jennewein
Research Computing Manager
Information Technology Services
The University of South Dakota
605.658.6068

_______________________________________________
Iplant-api-dev Mailing List: Iplant-api-dev at iplantcollaborative.org<mailto:Iplant-api-dev at iplantcollaborative.org>
List Info and Archives: http://mail.iplantcollaborative.org/mailman/listinfo/iplant-api-dev
One-click Unsubscribe: http://mail.iplantcollaborative.org/mailman/options/iplant-api-dev/rogerab%40email.arizona.edu?unsub=1&unsubconfirm=1
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.iplantcollaborative.org/pipermail/iplant-api-dev/attachments/20130809/815635ff/attachment.html