I'm also getting these types of errors in the log:
NOTICE: Deadlock detected -- See the lock(l) manual page for a
possible cause.
ERROR: WaitOnLock: error on wakeup - Aborting this transaction
NOTICE: Deadlock detected -- See the lock(l) manual page for a
possible cause.
ERROR: WaitOnLock: error on wakeup - Aborting this transaction
also some of these:
IpcSemaphoreCreate: semget failed (No space left on device)
key=5432614, num=16, permission=600
IpcSemaphoreCreate: semget failed (No space left on device)
key=5432714, num=16, permission=600
and many, many of these:
FATAL: pq_putnchar: fputc() failed: errno=32
FATAL: pq_putnchar: fputc() failed: errno=32
My disks are virtually empty and I can't see where I'm getting a
broken pipe.
All communications are local using DBI/DBD::Pg w/Perl 5.005_02.
Any thoughts? I'm thinking it might have to do with sockets/streams
since 7 uses sockets and all prior solaris 2.x used streams. Maybe
it's time to override some configure options?
DwD
-----Original Message-----
From: owner-pgsql-hackers@postgreSQL.org
[mailto:owner-pgsql-hackers@postgreSQL.org]On Behalf Of Tatsuo Ishii
Sent: Tuesday, February 09, 1999 11:01 PM
To: The Hermit Hacker
Cc: Daryl W. Dunbar; pgsql-hackers@postgreSQL.org
Subject: Re: [HACKERS] PostgreSQL and Solaris 7?
Reproduced here too. (Solaris 2.6)
>On Tue, 9 Feb 1999, Daryl W. Dunbar wrote:
>
>> Hello,
>>
>> I am running 6.4.2 on Sparc/Solaris 2.7. Occasionally (far more
>> often than I would like - which is never :), the postmaster
refuses
>> connections. Upon investigation, there are 48 postmaster
processes
>> running. If one kills the parent and restarts, all is well for
some
>> time. Attached is a chunk of output from postmaster,
unfortunately
>> it does not timestamp, so it is hard to see when what happened.
I'm
>> most disturbed about the device space message. None of my disks
are
>> even close to full, nor have been...
>>
>> Help! Advice? Ways to gather further info? Is this a Solaris 7
>> thing?
>
>We just got our Solaris 7 CDs, and I'm going to be installing it on
one of
>our quieter machines in coming weeks, but...
>
>have you tried doing a 'truss -p' on the parent process?
Increasing # of semaphores should solve the problem, I guess. I'm
going to try that as soon as I find the way to increase semaphores.
---
Tatsuo Ishii