Repository navigation
Deadline exceeded while reindexing all spaces #3237
Description
Describe the bug
running opencloud search index --all-spaces --insecure --force-rescan still gives me an rpc Error deadline exceeded.
Steps to reproduce
- Run command
opencloud search index --all-spaces --insecure --force-rescan - The error occures after some time.
Expected behavior
All spaces should be reindexed.
Actual behavior
Terminal throws an error:
[69/6001] indexed space b90f49a2-097c-42bc-83ff-278a76ab2e8b$6ff29e3e-8caf-4393-9f33-2be26b4ed8a8!6ff29e3e-8caf-4393-9f33-2be26b4ed8a8 in 42.570234898s
rpc error: code = DeadlineExceeded desc = context deadline exceeded
Setup
Opencloud rolling 7.4.0 on Debian 13 with Docker.
Additional context
I also tried to set SEARCH_REINDEX_CONCURRENCY=1 but that didn't fix the issue.
Activity
rhafer commentedon Aug 4, 2026
Yeah, that is what happens, even if the command line tool exits with that "deadline exceeded" error. The indexing process is happening on the server in the background AFAICT
BEBU88 commentedon Aug 4, 2026
I checked the CPU usage. It dropped from ~ 70 % to ~ 10% after the error occured. For me this indicates the indexing stopped but I will check this later.
BEBU88 commentedon Aug 5, 2026
I did some further testing and set the loglevel for the search service to debug while indexing. After indexing stopped I did a search for a specific word and the results where empty. In my opinion there should have been a result since the space should have been indexed then.
mwiedemeyer commentedon Sep 15, 2026
same after upgrading to v8
micbar commentedon Sep 15, 2026
Yes, same here.
Retriggering the reindex in a screen session, will see tomorrow if it succeeded.
Limerick-gh commentedon Sep 15, 2026
Got the same when doing rescan for v8.0.0 update.
Limerick-gh commentedon Sep 17, 2026
The deadline exceeded error is gone with the v8.0.1, therefore the reindex is eating up all memory and is being killed fter a while due to oom.
Adding extra memory did not fix it and I am getting the same result.
Is there a way to limit the memory usage? I already lowered concurrency to 1, but facing the same issue again.
aduffeck commentedon Sep 17, 2026
@Limerick-gh are you using a bleve or opensearch backend?
Is Tika enabled, and if yes is Tika or opencloud using the memory?
What amount of files/data and memory usage are we talking about?
What you can do as a workaround is running the command without the
--force-rescanbit which makes it pick up the work where it stopped last and not start all over again.