mirror of
https://github.com/fosrl/pangolin.git
synced 2026-07-16 06:08:33 -05:00
[GH-ISSUE #2800] Unable to upgrade from 1.16.2 to 1.17.0 - SQLite migration error #11005
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @cascer1 on GitHub (Apr 6, 2026).
Original GitHub issue: https://github.com/fosrl/pangolin/issues/2800
Originally assigned to: @oschwartz10612 on GitHub.
Describe the Bug
When starting Pangolin for the first time after upgrading from 1.16.2 to 1.17 it runs into a migration error.
Environment
To Reproduce
Expected Behavior
The migration should not fail
@karol-exe commented on GitHub (Apr 7, 2026):
I have the same problem
@LaurenceJJones commented on GitHub (Apr 7, 2026):
edit here a summary of all errors and I had to do manual sqlite fixes (I wont share them until maintainers investigate further)
debug info
1.17.0 SQLite Migration Issues
1. Missing
roleIdcolumn causes migration failureCause: Migration queries
roleIdfromuserOrgsanduserInvites(lines 17-32) without checking if the column exists.Fix: Check column existence before querying:
2. Users lose org access after migration
Symptom: Users can't access their orgs after upgrade.
Cause: Role assignments should migrate from
userOrgs.roleId→userOrgRolestable. IfroleIdcolumn doesn't exist,userOrgRolesstays empty and users have no roles.3. Runtime errors from missing tables
Cause: If migration fails early (due to number 1), tables created later in the transaction are never created.
@crisidev commented on GitHub (Apr 7, 2026):
I am having similar issues, specifically the first thing that fails for me is:
And then it trickles down to the same errors @LaurenceJJones shows in their debug:
I have reverted to 0.16.2 for now.
@oschwartz10612 commented on GitHub (Apr 7, 2026):
Hey @cascer1 or @crisidev do you have any more logs you could scroll up and send the whole chunk from where it says "starting migrations"?
@crisidev commented on GitHub (Apr 7, 2026):
I'll redeploy when I am home later and send the full logs, but the first sqlite exception is just after the line where it says migrations to version xyz done.
@crisidev commented on GitHub (Apr 8, 2026):
@oschwartz10612 full error logs after retrying to deploy 0.17.0:
@oschwartz10612 commented on GitHub (Apr 8, 2026):
Yeah this is because it has already failed the migration previously and
thinks it does not need to run it again. Do you have a backup of your
1.16.2 database? If so can you shut it down, load the old db, then
restart with the whole output?
@cascer1 commented on GitHub (Apr 9, 2026):
Here are the complete logs of the first startup of 1.17 using the 1.16.2 database.
When starting (using
docker compose up -d) it seems to hang indefinitely on the pangolin service. After one minute, I interrupted the process to collect the log. It looks like this output is generated in the first few seconds and after that nothing else is logged.https://gist.github.com/cascer1/ed7b9ac12d35e5559795f360e47119fc
@oschwartz10612 commented on GitHub (Apr 9, 2026):
Great thank you so much! This is what we saw in the discord thread as
well so now we have two pieces of evidence that this is a problem. I
will work on the patch.
@oschwartz10612 commented on GitHub (Apr 14, 2026):
This was adjusted I believe in the 1.17.1 so if you use that to go up from 1.16 it should work. Feel free to reopen if not.
@crisidev commented on GitHub (Apr 15, 2026):
I was able to upgrade from an old db on 1.16.2 and it worked properly with 1.17.1.
@sadorowo commented on GitHub (Apr 20, 2026):
In case someone who is using NixOS has already rebuilt their system and don't want to wait for a fix, here's a quick patch, add this to your overlays (assuming you have parameters named as
finalandprev)