pgBadger: Cannot find any log entries from systemd-journald
Hello, I think I just might did a small configuration issue, but somehow I cannot figure it out by myself. I want to let pgbadger directly fetch data from journald -- so I was using with a unexpected outcome: $ pgbadger --journalctl "journalctl -u postgresql.service" LOG: Ok, generating html report...s: 0, events: 0 Having a look into the journal there is a lot of ---cut Mar 05 09:06:23 myhost postgres[441820]: 2024-03-05 08:06:23 GMT [441820]: user=[unknown],db=[unknown],app=[unknown],client=xxx LOCATION: BackendInitialize, postmaster.c:4373 Mar 05 09:06:23 myhost postgres[441820]: 2024-03-05 08:06:23 GMT [441820]: user=[unknown],db=[unknown],app=[unknown],client=xxx LOG: 0: connection received: host=xxx port=34574 Mar 05 09:06:10 myhost postgres[441805]: 2024-03-05 08:06:10 GMT [441805]: user=postgres,db=xxx,app=psql,client=[local] LOCATION: log_disconnections, postgres.c:4722 Mar 05 09:06:10 myhost postgres[441805]: 2024-03-05 08:06:10 GMT [441805]: user=postgres,db=xxx,app=psql,client=[local] LOG: 0: disconnection: session time: 0:00:00.016 user=postgres database=xxx host=[local] Mar 05 09:06:10 myhost postgres[441805]: 2024-03-05 08:06:10 GMT [441805]: user=postgres,db=xxx,app=[unknown],client=[local] LOCATION: PerformAuthentication, postinit.c:292 Mar 05 09:06:10 myhost postgres[441805]: 2024-03-05 08:06:10 GMT [441805]: user=postgres,db=xxx,app=[unknown],client=[local] LOG: 0: connection authorized: user=postgres database=xxx application_name=psql ---cut--- My psotgresql.conf looks like: ---cut log_autovacuum_min_duration = 0 log_checkpoints = yes log_connections = yes log_destination = 'stderr' log_disconnections = yes log_error_verbosity = 'verbose' log_line_prefix = '%t [%p]: user=%u,db=%d,app=%a,client=%h ' log_lock_waits = yes log_min_duration_statement = 100 log_temp_files = 0 ---cut Any idea what I missed or did wrong? Cheers, Frank
Re: pgBadger: Cannot find any log entries from systemd-journald
Hello, Sorry for the late response. Am 06.03.24 um 16:40 schrieb Greg Sabino Mullane: On Tue, Mar 5, 2024 at 3:14 AM Frank Lanitz mailto:fr...@frank.uvena.de>> wrote: $ pgbadger --journalctl "journalctl -u postgresql.service" You could try adding --verbose to see if it gives more clues. No, unfortunately not. Though, it adds some more nice information in general, but none that actually helped me. Having a look into the journal there is a lot of None of the snippets from journald you showed were actually things pgbadger cares about, FWIW. It was just an random pick to show there is a bunch of logging -- I also can see checkpoints and vacuum etc. I just didn't want to paste 1GB of log ;) You can get an idea of what is actually parsed by running "journalctl -u postgresql --output=short-iso" Looks good to me. log_error_verbosity = 'verbose' This is probably not needed, and there is a finite chance that the extra context is confusing pgbadger. I tried to change it, but in my case no difference. Cheers, Frank
Re: Postgresql database terminates abruptly with too many open files error
Hello, Have you checked something like lsof to see open file descriptors to see? Cheers, frank Am 14.01.25 um 13:58 schrieb Sri Mrudula Attili: Hello Team, We have a postgresql VDB(virtual database- Delphix) that keeps terminating due "to too many open files". Below are few alerts that we could see from the postgresql.log < 2025-01-14 11:37:20.724 GMT >LOG: out of file descriptors: Too many open files in system; release and retry < 2025-01-14 11:37:20.724 GMT >FATAL: epoll_create1 failed: Too many open files in system The number of allowed openfiles at OS level are 65000. even then we see these all these being utilised and causing the database to terminate. Is there a way we could avoid this. Thanks, Sri
