Blitz Server restart notifications and too many open files?

[rpoehlmann]

Dear OME-Team and Community members,

we had a lot of "[OMERO] Server started" email notifications today (without actually triggering a manual service restart).
It seems like those notifications were cause by an automatically triggered restart of the Blitz service (see attached files "part1" and "part2"). Corresponding log entries at least seem to suggest this.

In addition, I could detect lots of the following errors in the "master.err" file (see attached file):

Code: Select all

!! 4/23/18 16:36:03:878 error: Blitz-0-Ice.ThreadPool.Server-19: fatal error: can't accept more connections:
   Too many open files
   0.0.0.0:43729


The user running OMERO has "ulimit unlimited".
Any ideas what might be causing those issues?

Any recommendations/feedback are greatly appreciated. Many thanks in advance for your help.

Best,
-Rainer

P.S.: I had to cut down the attached log-files, please let me know if you need additional information/output. Thanks!

[mtbc]

Dear Rainer,

All those "Too many open files" in master.err are certainly suggestive. Thank you for providing the logfiles, which I think suffice for this case. Just so you know, if it's more convenient in future do feel free to just zip them all up into one archive and submit to http://qa.openmicroscopy.org.uk/qa/upload/ (with an explanatory comment) though I do admit that it is nice to have the relevant parts extracted for us: when we get the full logs it isn't always even immediately clear which day and time to look for!

Would I be correct in suspecting that these problems arise from the very same system that is the topic of http://www.openmicroscopy.org/community ... f=4&t=8483 ?

Cheers,
Mark

[rpoehlmann]

Would I be correct in suspecting that these problems arise from the very same system that is the topic of http://www.openmicroscopy.org/community ... f=4&t=8483 ?

Yep, seems like we currently facing multiple issues at the same time, not clear yet if related or not ...
I will try to carefully read your reply regarding the other issue and then come back to you soon ...

Thanks a lot! Best,
-Rainer