Skip to content

Fix PostgreSQL invalid primary checkpoint record ΒΆ

This issue sometimes happens in the staging environment. The fix is to force the DB pod into a bash loop, while we run a shell command that will reset the database.

  • Configure the indevets-core-staging deployment:

    • Locate the definition for the indevets-core-staging-db container.

      command:
  - '/bin/bash'
  - '-c'
  - '--'
args:
  while true; do sleep 10; done;

    • Save the new configuration and wait for the new deployment to come online.

  • Access the indevets-core-staging-db pod shell and run the following commands:

    which pg_resetwal

Copy the output, it will be referenced in the next step as PG_RESETWAL_BIN_LOCATION

```bash
  su -l postgres
  PG_RESETWAL_BIN_LOCATION /var/lib/postgresql/data/
```

For example, the command could like: `/usr/lib/postgresql/11/bin/pg_resetwal /var/lib/postgresql/data/`