Post #1748145
2026-04-28 15:25 UTC
@phnt@fluffytail.org @m0xEE@breloma.m0xee.net @i@declin.eu doing god's work.
btw is the system maybe running out of ram and postgres is just OOM killed?
Replies (2)
-
@phnt@fluffytail.org 2026-04-28 15:33
@lain@lain.com @m0xEE@breloma.m0xee.net @i@declin.eu RHEL disables OOM on Postgres at least. Debian might not though. There was some runaway Erlang memory alloc issue, but I haven't seen that in a while and usually Pleroma is the one that gets OOMed.
-
@m0xEE@breloma.m0xee.net 2026-04-28 15:43
@lain@lain.com @i@declin.eu @phnt@fluffytail.org Unlikely, half of machine's RAM seems to be used as filesystem cache. Also I believe OOMkiller is reported to kernel log, but dmesg output seems clean. Timeouts due to silent storage failure seems a plausible theory, I'll check in a couple of hours if it returns and run a long SMART test.