Partial Summary (System Performance!!!!)

From: Awaneesh Mishra <sunsolin_at_yahoo.com>
Date: Tue Oct 09 2001 - 03:05:35 EDT
Techs !!

Sorry for the late summary.  A lot of replies but
somehow it did not help.  I am still waiting for oasis
in the desert...  Maybe some more replies will help...

Thanks

Awaneesh


Original Post: -
I have a system perfornace issue on a E450 box with 2
cpus and 1gb ram.  The system eats up around 66%
kernel resources. 


Responses: -

andy.lee@orange.co.uk 
Some thoughts : You are using disksuite , are you
using raid5 or mirror ? Is it only the rootdisk that
is mirrored , are you using the same
version/partitions disk suited on your 'equiv' test
box. (Raid5 can really suck kernel cpu) Have you got
the latest disksuite patches installed ? I suspect it
is down to disksuite. You could try mounting the
partition directly outside of disksuite control to
check . You seem to have high writes to md0 is your
app swapping a lot ? 


Trevor Paquette
Off the top of my head I am seeing ALOT of disk io to
some sort of meta device. You may want to off load all
that disk io to an actual REAL raid array, and not use
SDS. 


gez@beenz.com
Your md30 / 31 /32 seem to be a little high on the
svc_t is the Application only writing to these disks
and are there metadbs on them ? Can the load / metadbs
be distributed. 


Bertrand_Hutin
look at the results of sar -g pgscan/s should remains
>390 


Hindley Nick
Its a bit difficult to tell but you seem to have a lot
of I/O on md10 and from there onto c0t0d0. Do you have
a lot of errors being churned out into a log file
somewhere? 



__________________________________________________
Do You Yahoo!?
NEW from Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month.
http://geocities.yahoo.com/ps/info1
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
Received on Tue Oct 9 02:38:19 2001

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:42:26 EST