Pagecache size

[Marcelino Mata]

 I finally had a chance to rerun this large RAM job with the information
requested.  Attached are three files with the meminfo and sar reports.

You should notice that swapping is occurring while filesystem cache is
around 2-3GB RAM.

The only thing I have changed from the default in /etc/sysctl.conf is :


vm.bdflush=60 500 512 0 500 3000 87 50 0 
(bdflush default =  50      500     0       0       500     3000    80
50      0)

vm.kswapd=1024 32 16
(kswapd default = 512 32 8)

vm.pagecache=1 15 15 
(pagecache default = 1 15 30)

I just noticed that as I was replying to this post that I missed "vmstat
10 10" output.  I think the sar report has the information in less
detailed form.  I suppose vmstat would of given running tallies instead
of meminfo "snapshot in time" information.  If the vmstat information is
still required, let me know....

I have noticed that if I remove all swap partitions that my large job
stays in RAM during the run.  Not ideal solution since memory leaks will
come back and haunt me later.

Any suggestions would be very welcome,

Best Regards,

Marcelino

>-----Original Message-----
>From: redhat-list-bounces at redhat.com 
>[mailto:redhat-list-bounces at redhat.com] On Behalf Of Tolga Evren
>Sent: Thursday, October 13, 2005 1:18 AM
>To: General Red Hat Linux discussion list
>Subject: RE: Pagecache size
>
>Hi ,
>
>-Shouldn't you be using the hugemem kernel (or is that only for x86)?
>
>Yes, hugemem is only for x86 .
>
>Can you also provide the output of /proc/meminfo and sar -B 3 
>10000 , sar -W 3 10000.
>
>Kind Regards,
>Tolga
>
>
>
>-----Original Message-----
>From: redhat-list-bounces at redhat.com
>[mailto:redhat-list-bounces at redhat.com] On Behalf Of Ugo Bellavance
>Sent: Thursday, October 13, 2005 6:12 AM
>To: redhat-list at redhat.com
>Subject: Re: Pagecache size
>
>Marcelino Mata wrote:
>> Hi all,
>>  
>
>Hi
>
>> We have a Opteron workstation with 16GB RAM which we use to 
>run large 
>> memory jobs.  It seems that no matter what we set 
>> /proc/sys/vm/pagecache, the system starting swapping out after 12GB
>RAM
>> is used.  The largest job only needs about 13GB RAM so it 
>runs in RAM.
>> I have read the knowledge based docs and a few others but I 
>can't seem 
>> to address this issue.
>
>Can you provide us the output of 'vmstat 10 10' when memory 
>used > 12 GB?
>
>>  
>> RHEL 3.0 Update5  2.4.21-32.0.1.ELsmp (64bit)
>>  
>
>Shouldn't you be using the hugemem kernel (or is that only for x86)?
>
>> Marcelino
>>  
>>  
>>  
>>  
>>  
>
>
>--
>Ugo
>
>-> Please don't send a copy of your reply by e-mail.  I read the list.
>-> Please avoid top-posting, long signatures and HTML, and cut the
>irrelevant parts in your replies.
>
>-- 
>redhat-list mailing list
>unsubscribe mailto:redhat-list-request at redhat.com?subject=unsubscribe
>https://www.redhat.com/mailman/listinfo/redhat-list
>
>-- 
>redhat-list mailing list
>unsubscribe mailto:redhat-list-request at redhat.com?subject=unsubscribe
>https://www.redhat.com/mailman/listinfo/redhat-list
>
>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: sarW3-1000.txt
URL: <http://listman.redhat.com/archives/redhat-list/attachments/20051017/82d8cd22/attachment.txt>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: sarB3-1000.txt
URL: <http://listman.redhat.com/archives/redhat-list/attachments/20051017/82d8cd22/attachment-0001.txt>