public inbox for [email protected]  
help / color / mirror / Atom feed
Compiling postgresql
4+ messages / 3 participants
[nested] [flat]

* Compiling postgresql
@ 2003-03-06 18:16 Delao, Darryl W <[email protected]>
  2003-03-06 20:59 ` Re: Compiling postgresql Tom Lane <[email protected]>
  0 siblings, 1 reply; 4+ messages in thread

From: Delao, Darryl W @ 2003-03-06 18:16 UTC (permalink / raw)
  To: pgsql-docs

As it stands, I have postgres set to a max limit of 512 users.  I have been
told and read some material to where anything above 512 users may require re
compiling postgres.  I have installed postgres via a RPM and therefore have
never had to compile postgres before.  My main concern is what would I need
to change other than the user limit to make postgres to where it would
accept connections higher than 512 and what steps are involved in the
compiling process.  Right now, if I try to set the user limit above 512,
postgres simply will not start.  Any help is greatly appreciated.  
 
Thank you!
Darryl
 


^ permalink  raw  reply  [nested|flat] 4+ messages in thread

* Re: Compiling postgresql
  2003-03-06 18:16 Compiling postgresql Delao, Darryl W <[email protected]>
@ 2003-03-06 20:59 ` Tom Lane <[email protected]>
  0 siblings, 0 replies; 4+ messages in thread

From: Tom Lane @ 2003-03-06 20:59 UTC (permalink / raw)
  To: Delao, Darryl W <[email protected]>; +Cc: pgsql-docs

"Delao, Darryl W" <[email protected]> writes:
> never had to compile postgres before.  My main concern is what would I need
> to change other than the user limit to make postgres to where it would
> accept connections higher than 512 and what steps are involved in the
> compiling process.  Right now, if I try to set the user limit above 512,
> postgres simply will not start.  Any help is greatly appreciated.  

You mean 512 concurrent connections?  It sounds to me like you are
running into kernel-level limits on amount of shared memory or number
of semaphores.  But without seeing the error message the postmaster
puts out, it's hard to be sure.  See the discussion about "Managing
Kernel Resources" in the PG Administrator's Guide.

There has been no compile-time limit on number of connections for
several releases now, so I doubt you need to recompile.  (If you do,
you are overdue for an update.)

			regards, tom lane




^ permalink  raw  reply  [nested|flat] 4+ messages in thread

* Re: Compiling postgresql
@ 2003-03-06 21:02 Delao, Darryl W <[email protected]>
  2003-03-07 00:44 ` Re: Compiling postgresql Neil Conway <[email protected]>
  0 siblings, 1 reply; 4+ messages in thread

From: Delao, Darryl W @ 2003-03-06 21:02 UTC (permalink / raw)
  To: 'Tom Lane' <[email protected]>; Delao, Darryl W <[email protected]>; +Cc: pgsql-docs

Yes, 512 concurrent connections.  I have the db running on a dual xeon 2.4
ghz machine with 6 gig of ram...basically, im not even touching the ram at
this point.  If I wanted to increase to say 1024, then your saying I need to
look into the kernel-level limits?  When I tried to set it to 1024 before,
the postgres service wouldn't even start.  Do you think the reason for that
was some sort of kernel level problem or setting that needed to be tweaked?

Thanks,
Darryl



-----Original Message-----
From: Tom Lane [mailto:[email protected]] 
Sent: Thursday, March 06, 2003 2:59 PM
To: Delao, Darryl W
Cc: [email protected]
Subject: Re: [DOCS] Compiling postgresql

"Delao, Darryl W" <[email protected]> writes:
> never had to compile postgres before.  My main concern is what would I
need
> to change other than the user limit to make postgres to where it would
> accept connections higher than 512 and what steps are involved in the
> compiling process.  Right now, if I try to set the user limit above 512,
> postgres simply will not start.  Any help is greatly appreciated.  

You mean 512 concurrent connections?  It sounds to me like you are
running into kernel-level limits on amount of shared memory or number
of semaphores.  But without seeing the error message the postmaster
puts out, it's hard to be sure.  See the discussion about "Managing
Kernel Resources" in the PG Administrator's Guide.

There has been no compile-time limit on number of connections for
several releases now, so I doubt you need to recompile.  (If you do,
you are overdue for an update.)

			regards, tom lane



^ permalink  raw  reply  [nested|flat] 4+ messages in thread

* Re: Compiling postgresql
  2003-03-06 21:02 Re: Compiling postgresql Delao, Darryl W <[email protected]>
@ 2003-03-07 00:44 ` Neil Conway <[email protected]>
  0 siblings, 0 replies; 4+ messages in thread

From: Neil Conway @ 2003-03-07 00:44 UTC (permalink / raw)
  To: Delao, Darryl W <[email protected]>; +Cc: 'Tom Lane' <[email protected]>; pgsql-docs

On Thu, 2003-03-06 at 16:02, Delao, Darryl W wrote:
> Yes, 512 concurrent connections.  I have the db running on a dual xeon 2.4
> ghz machine with 6 gig of ram...basically, im not even touching the ram at
> this point.  If I wanted to increase to say 1024, then your saying I need to
> look into the kernel-level limits?

No, I think Tom is saying that you should provide the error message
you're encountering. For more information on configuring kernel-level
shared memory settings, see

http://www.ca.postgresql.org/users-lounge/docs/7.3/postgres/kernel-resources.html

Cheers,

Neil

P.S. User support should be done on lists like pgsql-general or
pgsql-novice, not pgsql-docs

-- 
Neil Conway <[email protected]> || PGP Key ID: DB3C29FC







^ permalink  raw  reply  [nested|flat] 4+ messages in thread


end of thread, other threads:[~2003-03-07 00:44 UTC | newest]

Thread overview: 4+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2003-03-06 18:16 Compiling postgresql Delao, Darryl W <[email protected]>
2003-03-06 20:59 ` Tom Lane <[email protected]>
2003-03-06 21:02 Re: Compiling postgresql Delao, Darryl W <[email protected]>
2003-03-07 00:44 ` Neil Conway <[email protected]>

This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox