[Bioclusters] requesting help for computational server setup

karthik viswanathan bioclusters@bioinformatics.org
Wed, 17 Sep 2003 17:13:04 -0500 (CDT)



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

yes i too noticed it, thats one of the reason i wanted to add 3 small hard
drives and put in an RAID 0, and move the current hard disks (146 GB each) to a
workstation and make it a file server.

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

adding one more cpu will definitely improve, but cost is a prob :-(

> 
> > 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....

Yes, I am not sure what is wrong and I am trying to troubleshoot, but the
current usage of the server by clients keeps me away from finding the problem. 


> 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.

Currently its striped 2. I thought this number should be equal to no of hard
disk and the raid will write files striping to these two hard disks. am i wrong?
kindly help me in figuring this out. One of the member in this group (Chris
Dagdigian) had also suggested to try XFS. I will definitely try it once i fix
with system configuration.

> 
> One thing to do, while others are running, is to use vmstat.  Run
> 
> 	vmstat 1
> 

$ /usr/bin/vmstat 1
   procs                      memory      swap          io     system      cpu
 r  b  w   swpd   free   buff  cache   si   so    bi    bo   in    cs us sy id
 2  0  0   1300  14084  86528 3414408    0    0    98   282  143   266 34  1 65
 1  0  0   1300  14404  86528 3412756    0    0   512    52  150   331 97  3  0
 1  0  0   1300  14404  86528 3412852    0    0   768     0  133   314 100  0  0
 1  0  0   1300  14404  86528 3412820    0    0   768     0  145   315 98  2  0
 1  0  0   1300  14404  86528 3412916    0    0   896     0  135   308 98  2  0
 1  0  0   1300  14404  86528 3412884    0    0   768     0  147   307 99  1  0
 1  0  0   1300  14404  86528 3412852    0    0   768    32  122   317 100  0  0

this was the output, will the swap be used only when the memory is excede the
system memory? I am not sure how the other numbers should look like.



> Joe
> -- 


Thanks
karthik