V18 -> V20 upgrade failed

[phales]

From reviewing the logs, it appear that during the upgrade process PostgreSQL is upgraded to version 13 but the script checking the version is expecting 11 and gets upset which causes the upgrade to fail.

Is this a known issue? If so, is there a timeline for a fix?

[33m====== [18:38:09] Preparation: Fetching current PostgreSQL version, requesting cluster version 11[39m
[33m====== [18:38:09] Preparation: Fetch postgresql information[39m
ii postgresql 13+225+deb11u1 all object-relational SQL database (supported version)
ii postgresql-13 13.16-0+deb11u1 amd64 The World's Most Advanced Open Source Relational Database
ii postgresql-client-13 13.16-0+deb11u1 amd64 front-end programs for PostgreSQL 13
ii postgresql-client-common 225+deb11u1 all manager for multiple PostgreSQL client versions
ii postgresql-common 225+deb11u1 all PostgreSQL database-cluster manager
[33m====== [18:38:09] Preparation: Fetch postgresql cluster: 13 main 5432 online postgres /var/lib/postgresql/13/main /var/log/postgresql/postgresql-13-main.log[39m
[33m====== [18:38:09] Preparation: Could not get postgresql cluster version from the settings database. Got: 13. Trying from pg_lsclusters result...[39m
[33m====== [18:38:09] Preparation: Unable to determine postgresql cluster version. Got: 13. Aborting...[39m
[33m====== [18:38:09] Upgrade Result: *******************************************[39m
{
"result": "aborted",
"time": "21-08-2024 18:38:09",
"title": "Unable to determine the PostgreSQL cluster version",
"message": "The script was unable to determine PostgreSQL cluster version. This may be a result of postgresql misconfiguration or service problem."
}
[31m[18:38] aborted: Unable to determine the PostgreSQL cluster version - 0[39m

I also noted the following in the logs:

Aug 21 18:36:55 lakeylaw cloud-init[265]: pkg_resources.DistributionNotFound: The 'cloud-init==23.4.4' distribution was not found and is required by the application

 

[bitn2]

Ist this pbx installed from iso and untouched or did you installed something else on the pbx? Did you started the upgrade form gui or console?

 

[Agathocles Prodromou]

Hi @phales ,

The upgrade initially failed due to a wrong Zabbix bookworm repository which caused it to abort midway. (we always ask users not to have third party repositories when upgrading as they are one of the most common reasons of failure)

Your system is now running Debian 11. Due to having already upgraded postgresql to 13, the script can no longer successfully proceed as it was not meant to "resume" when an upgrade fails midway. So no matter how many times you run it, you will always get the same error now. The behavior is expected.

You can either revert your system to Debian 10 as it was (if you took a snapshot before proceeding..), remove the Zabbix repository and run the upgrade again, or just spin up a new Debian 12/V20 instance using our ISO and restore from backup. The latter is a safer and probably faster way to migrate.

 

[phales]

We'll most likely spin up a new Debian v12/V20 and restore from backup.