Package: postgresql Version: 9.6+178 Severity: normal The default systemd unit file contains comments like
# restarting automatically will prevent "pg_ctlcluster ... stop" from working, [...] # (This should make pg_ctlcluster stop work, but doesn't:) I enabled Restart=on-failure and I am able to stop and start the cluster via systemd. Even on stable. So maybe this is outdated. == I enabled Restart=on-failure because 9.4+165+deb8u2 on stable, because it started to have problems with systemd communication on first start after system restarts, while working fine otherwise: systemd[1]: [email protected] start operation timed out. Terminating. So maybe it's not to bad to have a Restart=on-failure.
