From: | Daniel Jakots <postgres(at)chown(dot)me> |
---|---|
To: | pgsql-general(at)lists(dot)postgresql(dot)org |
Subject: | could not send data to client: Permission denied |
Date: | 2020-08-27 15:28:46 |
Message-ID: | 20200827111031.5ee46257@anegada |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Hi,
I run a flask rest api which use postgres to store data. I've been
using it for a few months and it worked flawlessly. I recently got an
error twice (58h10m apart). I added a lot more data a few days ago which
could be the reason I get this now, but the table is still quite small
with 69807 lines. This endpoint is called by a handful of machines
every 5 minutes and it works all the time except those two times.
The API runs in docker with gunicorn.
[2020-08-27 02:10:03,624] ERROR in app: Exception on /v1/pf [GET]
Traceback (most recent call last):
[...]
File "/home/snek/app/api/tools.py", line 94, in ip_get
cursor.execute("SELECT ip FROM pf_ip_ban;")
psycopg2.OperationalError: server closed the connection unexpectedly
This probably means the server terminated abnormally
before or while processing the request.
In postgresql log, it says:
Aug 27 02:10:02 db1 postgres[62464]: [7-1] 10.10.10.43(41816):[unknown](at)[unknown]: LOG: connection received: host=10.10.10.43 port=4181
Aug 27 02:10:02 db1 postgres[62464]: [8-1] 10.10.10.43(41816):api(at)api: LOG: connection authorized: user=api database=api
Aug 27 02:10:02 db1 postgres[62464]: [9-1] 10.10.10.43(41816):api(at)api: LOG: could not send data to client: Permission denied
Aug 27 02:10:02 db1 postgres[62464]: [9-2] 10.10.10.43(41816):api(at)api: STATEMENT: SELECT ip FROM pf_ip_ban;
Aug 27 02:10:02 db1 postgres[62464]: [10-1] 10.10.10.43(41816):api(at)api: FATAL: connection to client lost
Aug 27 02:10:02 db1 postgres[62464]: [10-2] 10.10.10.43(41816):api(at)api: STATEMENT: SELECT ip FROM pf_ip_ban;
Postgresql and the python code runs on two different VM, but
they're both on the same hypervisor (kvm on ubuntu). Therefore I would
assume the network between them should be uneventful; but software
has bugs.
Postgresql is postgresql-server-12.3p0 running OpenBSD -current.
I searched the error on the web but I didn't find anything with this
exact error. Do you know what could cause it? Or how could I get more
information?
I tuned OpenBSD according to our README:
https://github.com/openbsd/ports/blob/master/databases/postgresql/pkg/README-server#L59-L68
could it be that the values are too low since the table growth?
Cheers,
Daniel
From | Date | Subject | |
---|---|---|---|
Next Message | Fontana Daniel C (Desartec S.R.L.) | 2020-08-27 15:55:35 | Postgres and alias |
Previous Message | Tom Lane | 2020-08-27 13:44:33 | Re: Are advisory locks guaranteed to be First Come First Serve? And can the behavior be relied upon? |