Re: [ADMIN] problem starting 2 postmasters on sun solaris2.6

From: "Ross J(dot) Reedstrom" <reedstrm(at)wallace(dot)ece(dot)rice(dot)edu>
To: John Uhlig <juhlig(at)parc(dot)xerox(dot)com>
Cc: pgsql-admin(at)postgresql(dot)org
Subject: Re: [ADMIN] problem starting 2 postmasters on sun solaris2.6
Date: 2000-02-11 18:15:30
Message-ID: 20000211121530.A12448@rice.edu
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

Your probably better off with and actual Solaris expert answering this, but
I'll try to pass on the clues I've absorbed from this list:

On Fri, Feb 11, 2000 at 07:49:46AM -0800, John Uhlig wrote:
>
> Hi.
>
> I have spent most of a day trying to get 2 postmasters to run on a solaris2.6 sun.
> I have searched thru the admin mail archives but have not been able to find
> anything that has helped at this point.
>
> I am currently trying to run 2 postmasters (v6.5.3) one compiled with defaults
> and the other with a "-- with-pgport=5400" (arbitrary number). I am using
> a "-p 5400" option to start the 2nd postmaster and am setting the PGPORT
> and other related PG variables to keep the 2 postmasters separated.
>
> I am able to start one or the other separately but NOT both. I get error messages
> to the effect:
> ---------------------------------------------------
> FindExec: found "/opt/local/postgres-6.5.3_5400/bin/postgres" using PATH
> binding ShmemCreate(key=5265c1, size=1073152)
> IpcSemaphoreCreate: semget failed (No space left on device) key=5400015, num=16,

The postmaster tries to allocate all the shared memory buffers at startup, to avoid
running out in the middle of a large query.

> permission=600
> proc_exit(3) [#0]
> shmem_exit(3) [#0]
> exit(3)
> ------------------------------------------------------
>
> I have seen references to changes for /etc/systems file, but the few examples did
> not seem to work for me - perhaps because I did not have the right values?
>

The examples all assume one postmaster. You probably need to double some of those numbers.

Another possiblity: the error above is "No space left on device": That's
either the "can't get shared mem" error, or does Solaris allocate swap
space or something for all shared memory segments? Could it be you need
to add more swap?

Hope it helps,

Ross
--
Ross J. Reedstrom, Ph.D., <reedstrm(at)rice(dot)edu>
NSBRI Research Scientist/Programmer
Computer and Information Technology Institute
Rice University, 6100 S. Main St., Houston, TX 77005

In response to

Browse pgsql-admin by date

  From Date Subject
Next Message J Carlos Morales Duarte. 2000-02-11 20:56:58 Unlimited data or variable
Previous Message Bruce Momjian 2000-02-11 15:15:24 Re: [ADMIN] CLUSTER and SERAIL type