[Bioclusters] requesting help for computational server setup

Joe Landman bioclusters@bioinformatics.org
Wed, 17 Sep 2003 04:57:11 -0400


On Wed, 2003-09-17 at 16:41, karthik viswanathan wrote:
> Hi Joe:
> 
> Thanks for ur reply. 
> 
> " T Get 3 newhe performance ..." 
> this is a typo error, sorry about that. I am surprised how this got appended! it
> should have been
> "The performance is not satisfactory ..."

:)

> The programs the client run mostly are
> 
> 1. LUCY  (http://www.tigr.org/software/)

Somewhat disk intensive.

> 2. GENESEQER  (http://bioinformatics.iastate.edu/bioinformatics2go/gs/help.html)

More CPU intensive.

> I had ran hdparm
> 
> # /sbin/hdparm -tT /dev/sda2
>  
>  /dev/sda2: 
> Timing buffer-cache reads: 128 MB in 0.25  seconds =512.00 MB/sec 
> Timing buffered disk reads: 64 MB in 1.75 seconds = 36.57 MB/sec

Egad!  Thats low....

[root@squash landman]# hdparm -tT /dev/md0
 
/dev/md0:
 Timing buffer-cache reads:   128 MB in  0.30 seconds =426.67 MB/sec
 Timing buffered disk reads:  64 MB in  0.68 seconds = 94.12 MB/sec

These are two IDE drives in a software RAID0.

See if you can adjust the stripe unit on the hardware raid.  It would
require rebuilding the raided file system though.  Also, look at using
XFS rather than ext3.

One thing to do, while others are running, is to use vmstat.  Run

	vmstat 1

in a window, and watch the state of the machine.  Read the man page for
details on the fields.  

Joe
-- 
Joseph Landman, Ph.D
Scalable Informatics LLC,
email: landman@scalableinformatics.com
web  : http://scalableinformatics.com
phone: +1 734 612 4615