As the title suggests I’m very confused by what’s going on with Fly Postgres.
I’ve just tried to set up a new cluster and it appears that the entire stack for how fly Postgres is managed has been changed without warning or documentation?
It appears like stolon has been dropped in favour of repmgr – not a good move after investing time into getting familiar with stolon.
WAL-G support now appears to be missing from the image so now PITR exports don’t work.
What’s going on with this, how do I deploy a usable Postgres cluster now?
Well it certainly wasn’t communicated very well that the underlying tools have changed. But what about WALG? That doesn’t seem to be set up on the new images, which is required for a production database.
@LeoAdamek The move was made for reliability purposes, which is expressed in the post above. There are not many users who actually use WAL-G, so it would be unreasonable for us to hold back the transition for that reason alone. At the same time, we will continue to offer stolon as an option at provision time for those of you who rely on it. I hope that makes sense.
Ok understood, I thought wal-g would be much more widely used.
I guess for now I can manage snapshots using pg_dumpall. What’s the recommendation for PITR though, or is there simply nothing available to export WALs?