[Linux-cluster] newbie question about RHCS/GFS stability and performance

[Kamal Jain]

Hi,

Typical servers are Dell PowerEdge 2950s with single quad-core Intel X5365  @ 3.00GHz processors and 16GB RAM, onboard PERC 5/I RAID controllers, 146GB, 2.5", 10K-RPM SAS drives for /, /boot, and swap (32GB swap partition).

Using IOZONE v3.283 (detailed info below), with these types of commands:

$ ./iozone -Ra -g 2G -b ~/iozone.xls -f /mnt/gfsvol1/iozone/iozone.tmp

That runs the automatic matrix of IOZONE tests with test file sizes up to 2GB, and it skips small read/write requests for large files for time's sake.  It is running single thread, single I/O file.  I suspect our results will look somewhat different if I spin up more parallel threads and try to match IOZONE's testing thread-count to our application's.


The bottom line, however, is really around RHCS and GFS.  I'm hoping to find some good pointers for performance tuning, and monitoring and measurement to feed back into the tuning, best practices, etc.  There is a presentation at http://people.redhat.com/czinzili/docs/Summit07presentation_GFS_Best_Practices_Final.pdf which covers some things, but it's pretty light.  It did, however, point me at this mailing list, so there's a golden nugget!

Thanks,

- K






From: linux-cluster-bounces at redhat.com [mailto:linux-cluster-bounces at redhat.com] On Behalf Of Gary Romo
Sent: Wednesday, December 19, 2007 11:01 AM
To: linux clustering
Cc: linux-cluster at redhat.com; linux-cluster-bounces at redhat.com
Subject: Re: [Linux-cluster] newbie question about RHCS/GFS stability and performance


I will be using iozone at a later time.  Can you provide some syntax examples of what you are running?
Thanks.

Gary Romo
IBM Global Technology Services
303.458.4415
Email: garromo at us.ibm.com
Pager:1.877.552.9264
Text message: gromo at skytel.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20071219/70b336eb/attachment.htm>