PSQLException when uploading large datasets with Omero-Cli

Just an update, I tried with disabling Thumbnails only but it still fails. I assume it needs to disable all postprocessing.
Also on another note: I saw this trello and this is exactly what we see in the log: wondering if this is related to our fail:


Hi @Benjamin_Knorlein,

There’s not currently a way to skip solely post processing, but it occurs to me: is your processor running? Can you attach the output of:

omero admin diagnostics

and perhaps the contents of var/log/Processor-0.log?


Hi @joshmoore

Here the output from the diagnostics

OMERO Diagnostics (admin) 5.6.2

Commands:   java -version                  1.8.0 (/usr/bin/java)
Commands:   python -V                      3.6.8 (/opt/omero/server/venv3/bin/python -- 2 others)
Commands:   icegridnode --version          3.6.5 (/usr/bin/icegridnode)
Commands:   icegridadmin --version         3.6.5 (/usr/bin/icegridadmin)
Commands:   psql --version                 12.3 (/usr/bin/psql)
Commands:   openssl version                1.0.2 (/usr/bin/openssl)

Server:     icegridnode                    running
Server:     Blitz-0                        active (pid = 667, enabled)
Server:     DropBox                        active (pid = 677, enabled)
Server:     FileServer                     active (pid = 692, enabled)
Server:     Indexer-0                      active (pid = 696, enabled)
Server:     MonitorServer                  active (pid = 734, enabled)
Server:     OMERO.Glacier2                 active (pid = 749, enabled)
Server:     OMERO.IceStorm                 active (pid = 759, enabled)
Server:     PixelData-0                    active (pid = 776, enabled)
Server:     Processor-0                    active (pid = 793, enabled)
Server:     Tables-0                       active (pid = 808, enabled)
Server:     TestDropBox                    inactive (enabled) 

And here the Processor-0.log file Processor-0.log.txt (1.7 MB)

I also tested more uploads and it turns out that it works if i only disable the minmax computation with --skip minmax. I will run more tests today to confirm. But it seems the problem lies in the minmax computation and not the thumbnails. Do you have more details on this step. Do we actually need the minmax computation? It shows minmax values in the Webinterface so I am not sure what is actually computed during this step and if we can simply skip it.

And I also had a careful look at the data itself and found a difference between the small and big datasets. The small datasets do not use fields in the screen plate layout. The big datasets however have a maximum of 4 fields per well. For most of the wells all 4 fields are used but some wells do not use all 4 of them and show up in omero with empty images In this case it shows the minmax as -1/-1 and all parameters except the pixel size are set. I am wondering if this is could be related to the problem.

Best and Thanks,