Omero intermittent crashes and can't run batch image export

[Magic]

I've been tasked with taking an Omero installation at Stanford that was having problems while running on a Mac and getting it to be stable running on Linux.

Setting up the new Linux server and running Omero worked fine. The batch image export worked fine and the software seemed stable.

On the old mac there were two problems:
* Batch image export didn't work
* Every half hour or so the server would crash

I then did a backup and restore of the database from the Mac to the Linux server. I also copied over all the files in the Omero data dir.

That worked, but the same exact two problems followed me.

For batch image export instead of getting a page of options, we get the following message on both the Mac and Linux server...

Cannot load the script: 3406
Error: "No processor available! [0 response(s)]"
Please contact your administrator

(The 3406 is under Linux, on the Mac the number was a 4).

I've put the relevant log for what happens during the batch image export here: http://pastebin.com/G1pW9cd2

As for the random crashing every half hour or so. the log info is here: http://pastebin.com/19TVzU0c

I think the problem is something that followed us from the database or data directory (which again was setup by someone other than me on the Mac).

I'm hoping someone can tell me how to fix this.

The Linux server is Ubuntu 12.04.

Thanks so much for your help!

[jmoore]

"Can't find default group for root" sounds suspiciously like there's been some change in the root user (deactivation, removal from a group, etc.) Could you possible send me (off-list if you'd prefer), the contents of your experimenter, groupexperimentermap, and experimentergroup tables? Specifically I'm looking for the results of:

Code: Select all

omero=> select e.id as eid, e.omename as ename, l.id as link_id, g.id as gid, g.name as gname from experimenter e, groupexperimentermap l, experimentergroup g where e.id = l.child and l.parent = g.id and e.id = 0;

eid | ename | link_id | gid | gname 
-----+-------+---------+-----+--------
   0 | root  |       0 |   0 | system
   0 | root  |       1 |   1 | user
(2 rows)


Cheers,
~Josh

[Magic]

Josh...

Can you post or private message me your email address?

Thanks
Sam

[jmoore]

NB: already done.

[jmoore]

As a follow-up, the "omeName" value for experimenter:id=0 (i.e. root) looks to have been changed, which caused this issue. I've added a note to https://trac.openmicroscopy.org.uk/ome/ticket/10209 to also prevent this modification.

~Josh.