Closed ahkole closed 6 months ago
Not sure if related, but whenever I shutdown the AiiDA daemon I get a lot of these in the postgresql logs:
2022-11-02 21:21:27.638 CET [15338] ([unknown]) {aiida_qs_arnold_4f9e95979a19b7e84b704aaa3c5b9d4d : BdG_arnold_4f9e95979a19b7e84b704aaa3c5b9d4d} LOG: unexpected EOF on client connection with an open transaction
You can see in the log line that this is caused by a transaction with one of the AiiDA database tables. Maybe the connections aren't cleanly closed when stopping the daemon?
Heya, can you provide the procedure you used to initiate the archive; from the CLI or another way?
Heya, can you provide the procedure you used to initiate the archive; from the CLI or another way?
I created the archive using the CLI with
verdi -p profile1 archive create pseudos.aiida --groups pk
and then I tried to import it in a different profile using
verdi -p profile2 archive import pseudos.aiida
where pk
was the group pk of the pseudopotentials that was created by installing a set of pseudopotentials with aiida-pseudo
, profile1
was the name of the profile where I had installed the pseudos and profile2
the name of the profile where I wanted to copy the pseudos to.
sqlite3.ProgrammingError: SQLite objects created in a thread can only be used in that same thread. The object was created in thread id 140676517820224 and this is thread id 140676480427712.
I think its related to closing of the object store, similar to: https://github.com/aiidateam/aiida-core/issues/4899
sqlite3.ProgrammingError: SQLite objects created in a thread can only be used in that same thread. The object was created in thread id 140676517820224 and this is thread id 140676480427712.
I think its related to closing of the object store, similar to: #4899
Sounds like a very similar issue yes. The archive that was imported consisted of pseudopotentials, which have an associated file stored in the repository. So importing of the archive would require access to the object store.
I also see that that issue was listed as being fixed. Was that fix already included in v2.0.4. of aiida-core
? Could something similar be used to fix this issue?
Describe the bug
During the import of an archive I created I got the following exception:
It does report a success at the end of the import, and all the nodes in the archive seem to be present after the import, but this exception makes me a bit worried that maybe something is corrupted in the database now.
Is there anyway to check whether the profile's database is not corrupted?
Steps to reproduce
I don't know how to reproduce the error. This is the first time I get this exception during an import of an archive. I tried reproducing it by creating a clean new profile and importing the same archive there, but then the import completed successfully without any exceptions. I might try to delete the nodes that were imported and import them again in the same profile to see if it reproduces the error, but I haven't tried that yet.
Expected behavior
No sqlite exceptions during import
Your environment