#Facing issues in updating self hosted version v1.13.7 -> latest version

1 messages · Page 1 of 1 (latest)

foggy orchid
#

Hi Team,We’ve successfully provisioned a test environment on GCP (Compute Engine + Cloud SQL) and validated the infrastructure. While networking, IAP, and Nginx are all green ($/healthz$ returns 200), we are hitting a wall with the database migration.

Current Status:
Infrastructure: Verified. Networking and permissions are correct.
Stable Version: The app runs healthily on v1.13.7 with our imported production DB dump.
The Issue: When upgrading to latest or v1.16.4, the app fails due to schema mismatches.
Error Logs: We're seeing missing relations and columns, specifically relation "core.cronTrigger" does not exist and column ... standardId does not exist.
The Diagnosis: The UI reports "Unable to Reach Back-end" because the app logic fails on these missing elements, even though the transport layer is functional. It appears we cannot jump directly from v1.13.x to v1.16.x with our current data dump.

Our Plan: We are looking to implement a stepwise minor version upgrade path. Are there specific "stop-gap" versions or migration scripts we should prioritize to bridge the gap between v1.13.7 and the current release?

image_2.png
#

CC: @fringe sail

lavish spruce
#

Hellop @foggy orchid
could you please confirm you've followed general guidelines regarding incremental upgrade ?

#

Seems like you're missing cronTrigger database migration
This might be due to a cross version upgrade process
My reco here would be restoring your initial functional version and proceed to incremental upgrade
Please let me know !

#

Some recos

  • Always pull latest patch version of a given major.minor or directly consume the major.minor docker hub mirror ( v1.11 for example )
  • Clearing cache after an upgrade is required for 1.12 1.13 and is good practice if you encounter any regression just after an upgrade
  • Always check workspace database version after an upgrade, if it's not equal to the installed twenty version inspect twenty-server logs especially upgradeCommand for errors
foggy orchid
#

Thanks a lot @lavish spruce, we were indeed jumping versions which seemed to be the issue, we did an incremental upgrade after that and it seems to work fine. Thanks a lot for your suggestions.