Missing Admin panel v0.50.3

Hey there. As title states our admin panel is missing when upgrading to v0.50.3. Our database exsists since v0.12.0 and we have had a lot of issues since then. This issue is still going on https://discord.com/channels/1130383047699738754/1348588157662138421. Some extra info: Running in docker containers with Nixos modules Current twenty version v0.50.3 Nixos version 24.11 Postgres version 16.8 So my question is. What is the best solution to our situation? And is there a possibility we could fix our database?
No description
26 Replies
Caspersonn
CaspersonnOP4w ago
No description
Prastoin
Prastoin4w ago
Hello @Caspersonn, please double check your user has admin settings and impersonate access in db Seems like you missed an actor composite field type migration But unless I'm mistaken can't see a link between this errors and not having the admin panel displayed Have you upgraded minor by minor ? createdBy source and createdbyContext is a migration made either in 0.42 or 0.43 @Caspersonn when would you be available for a call today ?
Caspersonn
CaspersonnOP4w ago
Hey @Prastoin. My user has access in Database with admin rights. Are these migration commands documented somewhere? I have upgrade my version minor by minor but some upgrade had issues and to run the commands twice after a docker restart. I made backups for version v0.35.0 until v0.41.0, so i will probably try again today from version v0.41.0. Sure 🙃
Prastoin
Prastoin4w ago
Are these migration commands documented somewhere?
Yes here https://twenty.com/developers/section/self-hosting/upgrade-guide I think that's the best way let me know if you encounter any blocking point upgrading from your backup
Caspersonn
CaspersonnOP4w ago
Yeah. I used those commands to upgrade, right i am trying again maybe i made a user error. I'l let you know if it worked!
Prastoin
Prastoin4w ago
Sounds good !
Caspersonn
CaspersonnOP4w ago
Okay @Prastoin. I encounterd my first issue when upgrading from v0.40.3 to v0.41.5. Most times i run both upgrade commands and restart and run them again. But is that the correct way?
No description
Prastoin
Prastoin4w ago
Why would you run them twice ?
Caspersonn
CaspersonnOP4w ago
Like this:
docker exec -it twenty-worker yarn command:prod upgrade-0.41
docker exec -it twenty-worker yarn database:migrate:prod

systemctl restart docker-twenty-worker.service docker-twenty.service docker-redis.service

docker exec -it twenty-worker yarn command:prod upgrade-0.41
docker exec -it twenty-worker yarn database:migrate:prod
docker exec -it twenty-worker yarn command:prod upgrade-0.41
docker exec -it twenty-worker yarn database:migrate:prod

systemctl restart docker-twenty-worker.service docker-twenty.service docker-redis.service

docker exec -it twenty-worker yarn command:prod upgrade-0.41
docker exec -it twenty-worker yarn database:migrate:prod
Don't know. It thought it would build the tables and such
Prastoin
Prastoin4w ago
I can't attest that those old commands are 100% idempotent tbh One should be enough Have you had any errors at the very first upgrade run ?
Caspersonn
CaspersonnOP4w ago
Yes
Prastoin
Prastoin4w ago
The ones in the screenshot ?
Caspersonn
CaspersonnOP4w ago
Jep, exactly the same i think
Prastoin
Prastoin4w ago
I have to finish few things before, checking in an hour. Will have a look to mcmas command definition and also would be looking for hostname migration in anterior migration 🤔
Caspersonn
CaspersonnOP4w ago
Alright. I am open to call, if you prefer that? 😄
Prastoin
Prastoin4w ago
Yes lets setup a call for 4pm Paris TZ if you're also available, please let me know
Caspersonn
CaspersonnOP4w ago
Sure, that sounds good to me!
Caspersonn
CaspersonnOP4w ago
Hey @prastoin. I have done the migration to a new workspace and it was successful for the most part, there are only a few issues. The admin panel is still missing somehow, and i got 59 issues from yarn command:prod workspace:health -w {workspaceId}. Are you open for another call?
No description
Caspersonn
CaspersonnOP4w ago
Ooh and also have these issues when opening some companies. I have upgraded to v0.50.9 as well with no errors.
No description
No description
Caspersonn
CaspersonnOP4w ago
When create a new "clean" database with version v0.50.9 with the command yarn command:init:prod, i get exactly the same amount of issues. Is this normal?
No description
Caspersonn
CaspersonnOP4w ago
Found out that this has something to do with the timelineActivity i imported. When a company, Opportunity or people has a note or task from my database than it just doesn't work. When i remove those specific linked notes or tasks in the table timelineActivity than it works.
Prastoin
Prastoin3w ago
Hello @Caspersonn ! Sorry for the delay Glad to see you find your way through in migrating your instance to twenty latest version ! Regarding your workspace healthcheck issues it seems to come from the data your imported The actor composite especially, should have a context defaulted value to an empty object etc You should be able to massively fix them quite easily Regarding this specifically I'm not sure this seems to be some "should never occur invariant assertion to me* could it be a warmup cache issue or invalid data in database ? @charles what do you opt for ?
Caspersonn
CaspersonnOP3w ago
Hey @Prastoin. No worries! The issues where also there when i created a new blank workspace, so that seems kinda odd. If you want a better look in to how i did the migration, i would be open for another call!
Prastoin
Prastoin3w ago
Ah ! Maybe a known bug, seems to be workspace agnostical then Would love to see how you have done ! When would you be available ?
Caspersonn
CaspersonnOP3w ago
Okay. I am available Today, between 11 am and 12 am Paris time or later this afternoon.
Prastoin
Prastoin3w ago
Dming you to setup the call

Did you find this page helpful?