Alex Simenduev

> I'd also be interested to understand the purpose behind changing the service name. With the recent removal of multi-instance support in the 7.x module release, we're trying to stick...

That's what I thought too, but seem like it isn't. `[notice]` logs only from `mruby` module are printed, for example `perl` module doesn't print anything , see config below ......

@floreks the second part never works, when it fails back to `/bin/sh` the console closes right after shell started.  If `/bin/bash` present, everything works well!

@Ormod nope, the server fails to start ``` 2017-07-04 11:21:06 UTC [8914-1] pghoard@[unknown] FATAL: the database system is starting up 2017-07-04 11:21:06 UTC [8915-1] pghoard@[unknown] FATAL: the database system is...

@Ormod here are the logs, after each step (https://gist.github.com/shamil/6386042b222906c338382e8b2340da3c) Thanks

interestingly enough I tried similar scenario with `wal-e` and everything worked out of the box. Long term `wal-e` is not an option for me, because it doesn't support backups from...

Tried that, even invoked `select pg_switch_xlog();` to trigger wal switch. Similar problem.

In the `base_label` file I see that `START WAL LOCATION` is set to `0000001E0000000200000052` but `pghoard` started to upload WAL files from `0000001E0000000200000054` which basically means we have 2 missing...

`gcp_compute_disk_info` returns a list of matched disks, in your case it will have only one element, so I guess changing the `gcp_compute_snapshot` task to below can help: ```yaml - name:...

Maybe having an optional param in `EncryptedSecret` resource to force recreate the secret, this can help and let people decide what to do in such cases;