[Clas12_software] osg server possible issues

[Maurizio Ungaro]

Dear all,

The OSG server has been occasionally struggling to keep up with I/O demands.

Thomas Britton also noticed that there were GLUEX jobs which completed successfully with exit code = 0 but did not produce data.  The culprit seems to be xrootd.
xrootd is used by GLUEX to transfer data to the farm nodes.

CC had a meeting and came up with a plan to offload xrootd (this was intended in any case) from the osg server.
xrootd data is now being transferred to another server. This move is adding load on the osg server.

We will keep monitoring jobs closely to make sure all data is transferred back at JLAB.

Regards,

Mauri


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.jlab.org/pipermail/clas12_software/attachments/20191122/67c65b9a/attachment.html>