|
|
sybperl-l Archive
Up Prev Next
From: Michael Burstin <mikeb at concerto dot com>
Subject: Re: (OT) Sybase memory footprint
Date: Jan 13 2003 2:15PM
Well, without seeing the error, I can guess it could be one of a few
things.
1. There is a .krg file in $SYBASE.
2. You have too low of a value of shminfo_shmmax value in your
/etc/system file.
Look for a line saying:
set shmsys:shminfo_shmmax=
in your /etc/system file. The value for this must be >= your Sybase
memory size. This value is set in bytes, so take 384 * 1024 * 1024
for your value. This value requires a reboot to take place.
On Mon, Jan 13, 2003 at 03:03:09PM +0100, David Landgren wrote:
> Hello,
>
> First and foremost, happy new year and sorry for this slightly off-topic
> question.
>
> I have a Sybase 11 installation running on a Solaris box with 512Mb. It
> is configured so as to give Sybase exactly half the RAM, i.e. 256. I
> recently upgraded the machine to 1Gb RAM.
>
> I would like to tell Sybase to use 384Mb now. As things stands, the
> config file (hostname.cfg) has all values = DEFAULT, apart from the
> following:
>
> cache size = 10M
> cache status = log only
> pool size = 9.0000M
> cache status = default data cache
> total memory = 128000
> number of user connections = 64
> number of locks = 12500
>
> I was under the impression that I just had to scale up the "total
> memory" setting. When I changed it to 192000 and then 144000 and even
> 130000, Sybase refused to start and emitted a kernel / vmem kind of
> error. No, I didn't write the error down, I was nervous and more
> interested in just having Sybase run, at any amount of memory used. So I
> returned to the initial value and everything is fine.
>
> So the question is what did I do wrong, or what did I not do that I
> should have?
>
> Thanks,
> David "formerly dlandgre@bpinet.com" Landgren
>
--
Michael Burstin
NPI Engineer
Concerto Software
978-952-0842
mikeb@concerto.com
|
