Verified Commit cc2d91b2 authored by anarcat's avatar anarcat
Browse files

disable backups on Prometheus WAL files (team#41627)

Normal chunks are okay to backup without any special snapshotting, but
WAL files are written on the fly by Prometheus, and doing a backup
with Bacula will lead to inconsistencies in the storage which can make
recovery more difficult. See also https://github.com/prometheus/prometheus/pull/14317
parent ecc7fd75
Loading
Loading
Loading
Loading
+5 −2
Original line number Diff line number Diff line
@@ -952,8 +952,11 @@ require little backups. The metrics themselves are kept in
`/var/lib/prometheus2` and should be backed up along with our regular
[backup procedures](howto/backup).

This might be incorrect: upstream documentation says we need special
procedures for backups to be consistent, see [tpo/tpa/team#41627](https://gitlab.torproject.org/tpo/tpa/team/-/issues/41627).
WAL (write-ahead log) files are ignored by the backups, which can lead
to an extra 2-3 hours of data loss since the last backup in the case
of a total failure, see [tpo/tpa/team#41627](https://gitlab.torproject.org/tpo/tpa/team/-/issues/41627) for the
discussion. This should eventually be mitigated by a high availability
setup ([tpo/tpa/team#41643](https://gitlab.torproject.org/tpo/tpa/team/-/issues/41643)).

## Other documentation