Hi there,
I’ve created a new database, which was in fra on two volumes.
I’ve created then another volume in cdg, deleted one of the initial fra volumes and restarted the app.
Now I keep getting tons of messages in the logs:
2021-05-08T13:13:15.024Z app[fa843eae] fra [info] sentinel | 2021-05-08T13:13:15.019Z WARN cmd/sentinel.go:276 no keeper info available {"db": "2fbe9e2c", "keeper": "fdaa02752a7b6701d6e2"}
Even thought the status, replicas and checks all pass. Is this normal? Is there documentation on how to set up the HA of a PostgreSQL cluster other than how I did it?
Ah this is a confusing part of our postgres interface.
The “right” way to add regional replicas is to create a new volume in the region you want, flyctl scale count 3, and then wait. Once flyctl status shows that it’s a healthy follower, you can flyctl volumes remove <old-volume> and then flyctl scale 2.
It seems like that db has been deleted, if you feel like giving it another try with a new one please let us know how it goes!
I followed your steps, and I keep getting the warnings.
Besides that, scaling down to 2, didn’t pick the new instance from cdg, but kept the old one with the deleted volume. I scaled the app down to 1 and back to 2 to get the cdg replica up, thought with the warning message:
2021-05-10T17:32:30.705Z app[f19ed761] fra [info] sentinel | 2021-05-10T17:32:30.703Z WARN cmd/sentinel.go:276 no keeper info available {"db": "8839a357", "keeper": "fdaa02752a7b6601df72"}