Bug: "No Processor Available"

[wmoore]

There's been some investigation of this, so let's put everything we know in one place...

nightshade was suffering from this error for all scripts, E.g.

• ​http://qa.openmicroscopy.org.uk/qa/feedback/4242/
• ​http://qa.openmicroscopy.org.uk/qa/feedback/4237/

devteam:

looks like it's completely dead in the water:
2012-02-09 20:44:24,290 INFO  [                       omero.util.Server] (MainThread) ********************************************************************************
2012-02-09 20:44:24,297 INFO  [                       omero.util.Server] (MainThread) Starting
2012-02-09 20:44:24,415 INFO  [                    omero.util.Resources] (Thread-2  ) Starting
2012-02-09 20:45:22,610 INFO  [                             omero.utils] (MainThread) Failed to get session on attempt 1
2012-02-09 20:45:32,612 INFO  [                             omero.utils] (MainThread) Failed to get session on attempt 2
2012-02-09 20:45:42,790 INFO  [                             omero.utils] (MainThread) Failed to get session on attempt 3
2012-02-09 20:45:52,792 INFO  [                             omero.utils] (MainThread) Failed to get session on attempt 4
2012-02-09 20:46:02,794 INFO  [                             omero.utils] (MainThread) Failed to get session on attempt 5
2012-02-09 20:46:12,913 INFO  [              omero.processor.ProcessorI] (MainThread) Registering processor Processor -t @ Processor-0.ProcessorAdapter
2012-02-09 20:46:12,946 INFO  [                       omero.util.Server] (MainThread) Entering main loop

wait that's february.

This was fixed with

$ bin/omero admin ice "server stop Processor-0"

Possible workaround: add a maxExecutions parameter so that after a number of jobs the process restarts (via icegridnode). This would possible require some locking, but if 2 processors were running, this would likely not be a problem. This would only until processor.py is replaced with an MQ #7902.

[wmoore]

More of the same

• ​https://www.openmicroscopy.org/community/viewtopic.php?f=4&t=1230&p=4830
• ​http://qa.openmicroscopy.org.uk/qa/feedback/4498/
• ​http://qa.openmicroscopy.org.uk/qa/feedback/4499/

[jmoore]

I've opened a PR that may have a little effect on this issue: ​https://github.com/openmicroscopy/openmicroscopy/pull/220

It's unlikely to make it go away, but at least will have more information going forward.

My suspicion is that a lot of the issues are caused by network issues such that the main server is simply not able to contact the processor service. If that's the case, there's little we can do without a major re-architecting.