Re: [HACKERS] pq_recvbuf: unexpected EOF on client connection - Mailing list pgsql-hackers

From Tom Lane
Subject Re: [HACKERS] pq_recvbuf: unexpected EOF on client connection
Date
Msg-id 18562.940546074@sss.pgh.pa.us
Whole thread Raw
In response to pq_recvbuf: unexpected EOF on client connection  (Oleg Bartunov <oleg@sai.msu.su>)
List pgsql-hackers
Oleg Bartunov <oleg@sai.msu.su> writes:
> I notice  messages like:
> pq_recvbuf: unexpected EOF on client connection
> What does it means ?

Means your client closed the connection without sending a "terminate"
message first, ie, you didn't close down libpq gracefully.  It's harmless
enough, although I think having the log message is a good idea.  (If you
use clients that are careful to do PQfinish() then you can use the
postmaster log to check for client crashes.)

> Also I'm curious about postmaster's activity:

> proc_exit(0) [#0]
> shmem_exit(0) [#0]
> exit(0)
> /usr/local/pgsql/bin/postmaster: reaping dead processes...
> /usr/local/pgsql/bin/postmaster: CleanupProc: pid 21507 exited with status 0

> This message appears too often - I have in httpd.conf
> MaxRequestsPerChild 5000, so I expect new httpd children after 5000 requests
> and new postgress process accordingly (I use persistent connection
> between httpd and postgres).

Well, that's certainly the trace of a backend quitting.  I'd say your
httpd stuff isn't working the way you think it is...
        regards, tom lane

PS: I didn't hear back from you about INTERSECT/LIMIT --- is that still
broken for you?  I can't find anything wrong with it here.


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: [HACKERS] Neverending query on 6.5.2 over Solaris 2.5.1
Next
From: Vadim Mikheev
Date:
Subject: Re: [HACKERS] Planning final assault on query length limits