[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Problems with swapping



I have similar woes.. In my _opinion_, the problem is related to RHES
being too aggressive with caching when a lot of disk IO is present. We
have Oracle running and the system is in a constant state of swap even
though there is tons of cache memory available.

See these:

https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=118706 
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=124576 

I am waiting on a possible patch....but this is _my_ problem, you might
be experiencing a different issue..

Kevin.

P.S. I tried these settings for a while and it actually made my system
unstable (worse than swapping :-)

vm.bdflush = 10 1000 500 5000 0 6000 100 0 0 
vm.overcommit_memory = 0 
vm.overcommit_ratio  = 75 
vm.pagecache = 1       10      10 



On Fri, 2004-07-02 at 05:37, Mindaugas Riauba wrote:
> > >   Why to swap in/out when there is >3GB of data in cache? The single
> running
> > > task
> > > is Oracle. sysctl settings are default.
> > >
> > >   What could I tune or where to read about reducing swapping?
> >
> > http://people.redhat.com/nhorman/papers/papers.html
> 
>   Not much. From what I understood I tried to change vm.pagecache from
> "1 15 100" to "1 15 50" but it does not seem to help much.
> 
>   I found (http://openacs.org/forums/message-view?message_id=190696) that
> on SLES there is sysctl vm.vm_mapped_ratio which seems to be directly
> influencing
> swappiness. But there is no such sysctl on RHES.
> 
>   Mindaugas
> 
> 
> --
> Taroon-list mailing list
> Taroon-list redhat com
> http://www.redhat.com/mailman/listinfo/taroon-list



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]