[ome-devel] Issue after upgrading from 4.4.5 to 4.4.9

[Josh Moore]

On Nov 15, 2013, at 4:51 PM, Niko Ehrenfeuchter wrote:

> Hey Josh,
> 
> On 14.11.2013 17:26, Josh Moore wrote:
> [...]
>> 
>> After getting a sneak peak at your DB (thanks for that!), I think I have the answer:
>> 
>>  $ bin/omero hql "select g.id from ExperimenterGroup g join g.groupExperimenterMap m join m.child as e where e.omeName = 'diogo'"
>>  # | Col1
>> ---+------
>>  0 | 353
>>  1 | 1
>> (2 rows)
>> 
>> In other words, user "diogo" is no longer a member of group 3. How that happened, of course, I don't know. (Likely detectable in the Blitz-0.logs if you'd like)
> 
> the one thing that puzzles me now is that group 3 is the "default" group. The images should rather have been assigned to group 353, am I wrong here?

That depends on how "diogo" was logged in and what UI choices were made. (Unless the import happened via the CLI)


> Digging a bit around, I found some more users whose images were assigned to group 3 (whereas for most users their images are assigned to a group ID that makes sense in my feeling).
> 
> As it wasn't me setting up the group or user account in OMERO, I have no clue what was done there. However, this is the group membership of all the users that have images being assigned to group 3:
> 
>>> omero=> SELECT * FROM groupexperimentermap WHERE child IN (SELECT DISTINCT owner_id FROM image WHERE group_id = '3');
>>> id  | permissions | owner | version | child | external_id | parent | child_index
>>> -----+-------------+-------+---------+-------+-------------+--------+-------------
>>> 555 |        -120 | f     |       0 |   552 |             |    353 |           0
>>> ...
>>> 403 |        -120 | f     |       0 |     4 |             |    103 |           4
>>> 510 |        -120 | t     |       0 |     4 |             |    354 |           5
> 
> -> as you can see, none of these users is member of group 3...?!
> 
> Currently I don't really think so, but might this be related to our old "default group" issue?
> 
> http://www.openmicroscopy.org/community/viewtopic.php?f=5&t=1140
> 
> One more thing that confuses me even more: *my* user account is member of both relevant groups (353 and 3), yet I don't see the images either.
> 
>>> omero=> SELECT * FROM groupexperimentermap WHERE child IN (SELECT id FROM experimenter WHERE omename = 'ehrenfeu');
>>> id  | permissions | owner | version | child | external_id | parent | child_index
>>> -----+-------------+-------+---------+-------+-------------+--------+-------------
>>>  16 |        -120 | f     |       0 |     9 |             |      1 |           1
>>> ...
>>>  17 |        -120 | f     |       0 |     9 |             |      4 |           4
>>> 603 |        -120 | f     |       0 |     9 |             |    353 |           0
>>> (5 rows)
> 
> So I'm tempted to believe this is more something related to the upgrade, no? Should I upload the Blitz-0.logs? I could also offer a dump of the DB as it as with 4.4.5, if that helps?

The upgrade wouldn't haven't modified the DB tables at all. We can compare your before and after dumps to be absolutely certain. More likely is that some change was made in the group structure. My assumption would be once all the users have been added to the appropriate groups, everything should be visible when you log into that group for viewing. After that, they can log in and move their data to the appropriate group.

Alternatively, if you know where the data should go, we can work to do it via the CLI. Just let us know.

> Cheers
> ~Niko


Cheers,
~Josh