pg.ddx.io pgsql-admin@postgresql.org mailing list archive
help / color / mirror / Atom feedSSL broken pipe
2+ messages / 1 participants
[nested] [flat]
* SSL broken pipe
@ 2007-03-28 12:43 Marc Cousin <mcousin@sigma.fr>
2007-04-02 09:39 ` Re: SSL broken pipe Marc Cousin <mcousin@sigma.fr>
0 siblings, 1 reply; 2+ messages in thread
From: Marc Cousin @ 2007-03-28 12:43 UTC (permalink / raw)
To: pgsql-admin
Hi,
I've been having this problem occurring a few times in the last few days : in
the postgresql log, I've seen these messages :
2007-03-28 13:53:34.845 CEST,18249 LOG: SSL SYSCALL error: Broken pipe
2007-03-28 13:53:34.846 CEST,18249 LOG: SSL SYSCALL error: Broken pipe
2007-03-28 13:53:34.846 CEST,18249 LOG: SSL SYSCALL error: Broken pipe
2007-03-28 13:53:34.846 CEST,18249 LOG: SSL SYSCALL error: Broken pipe
2007-03-28 13:53:34.847 CEST,18249 LOG: SSL SYSCALL error: Broken pipe
2007-03-28 13:53:34.847 CEST,18249 LOG: SSL SYSCALL error: Broken pipe
2007-03-28 13:53:34.847 CEST,18249 LOG: SSL SYSCALL error: Broken pipe
2007-03-28 13:53:34.848 CEST,18249 LOG: SSL SYSCALL error: Broken pipe
2007-03-28 13:53:34.848 CEST,18249 LOG: SSL SYSCALL error: Broken pipe
As you can see, these messages are sent quite fast, and fill the filesystem
quite rapidly.
The server is postgresql 8.2.3 64 bits on debian unstable
The client is pgadmin3 as provided with Postgresql 8.1.7 for Windows.
I don't know what the client was doing exactly but I could investigate.
Is this a known problem ? It's quite problematic, as it can fill a gigabyte of
log file in a minute...
Thanks in advance
Marc
^ permalink raw reply [nested|flat] 2+ messages in thread
* Re: SSL broken pipe
2007-03-28 12:43 SSL broken pipe Marc Cousin <mcousin@sigma.fr>
@ 2007-04-02 09:39 ` Marc Cousin <mcousin@sigma.fr>
0 siblings, 0 replies; 2+ messages in thread
From: Marc Cousin @ 2007-04-02 09:39 UTC (permalink / raw)
To: ; +Cc: pgsql-admin
It seems to be what occurred : the windows client crashed ...
Still, it's a bit of a problem to have a filesystem full in a matter of
minutes just because a client machine crashed while having an SSL connection
opened on the postgresql server...
On Thursday 29 March 2007 19:40:18 Peter Koczan wrote:
> I saw this on one of my servers a few weeks ago (it was running
> postgres 7.4under linux, and I believe the users in question were just
> running psql). I
> talked to someone more knowledgeable about SSL about this, and he said that
> these errors are usually due to some abrupt disconnection somewhere (in any
> application, not just postgres). My best guess is that the connection was
> broken while data was trying to be fed, causing the server to say "Hey,
> where'd you go?...Here's your data again...Hey where'd you go?..." ad
> infinitum.
>
> Other than that, I have no idea what would cause this or how to fix it
> short of restarting the postgres server or rebooting the machine.
>
> That being said, I don't know a whole lot about SSL, and people with more
> knowledge are more than welcome to weigh in.
>
> Peter
>
^ permalink raw reply [nested|flat] 2+ messages in thread
end of thread, other threads:[~2007-04-02 09:39 UTC | newest]
Thread overview: 2+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2007-03-28 12:43 SSL broken pipe Marc Cousin <mcousin@sigma.fr>
2007-04-02 09:39 ` Marc Cousin <mcousin@sigma.fr>
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox