Postgres upgrade docker reports "already at version 11"

Technical Help
, ,
1

I’m trying to upgrade from 1.22.1 to 1.24.2 and started by running the pgautoupdate docker. I use portainer.io to manage containers so I stopped the main baserow container and duplicated the container by renaming and pulling baserow-pgautoupdate:1.24.2. I checked the logs after it exited and got this:

image
image796×505 15.6 KB

Which seems odd, I checked the PG_VERSION in the postgres folder and it does say 11. Everything in the folder seems untouched compared to a back up tar. For now, I have restarted my old container and everything seems to be working.

I’m not sure what the script is looking for to verify the upgrade, but shouldn’t it be looking for 15?

I have verify this behavior by restoring a backup to a test directory and running the upgrade docker on the test, generates the same message. Is there maybe an env that’s affecting this? I noticed the container has these env:

POSTGRES_BIN_FOLDER	/usr/lib/postgresql/15/bin
POSTGRES_LOCATION	/etc/postgresql/11/main
POSTGRES_OLD_BIN_FOLDER	/usr/lib/postgresql/11/bin
POSTGRES_OLD_LOCATION	/etc/postgresql/11/main
POSTGRES_OLD_VERSION	11
POSTGRES_SETUP_SCRIPT_COMMAND	upgrade
POSTGRES_VERSION	11
2

I have determined the issue is related to duplicating the baserow settings in portainer.io. The tool duplicates many of the default evn variables that comes with the previous docker image, and in turn causes problems when upgrading. The Postgres version and related env caused issues with both the database upgrade image and later on with the 1.24.2 image.

I was over zealous in following the upgrade instructions to include “ALL THE ARGUMENTS YOU NORMALLY ADD TO YOUR BASEROW INSTANCE”. When in fact, the database upgrade just needs access to the baserow_data directory and nothing else. And the main 1.24.2 image required that I remove old env that carried over from the last image.