UNIX for Advanced & Expert Users

Hi.

my company's mailbox server has severe performance issue on the days while full backup is running. There is huge I/O in another "mail replication process" for standby servers ( which runs all the time, every 15 mins. ).

specs of servers :- RHEL5.5 , kernel -2.6.18-53.1.6.el5 , 8 core Intel Xeon 2.66 GHz, 32 GB RAM.24 GB VM.

strangely I see

swappiness =0 ( proposing it to be 50 )

HugePages_Total: 0 ( proposing it to be 100 )
HugePages_Free: 0
HugePages_Rsvd: 0
Hugepagesize: 2048 kB ( proposing it to be 4096 )


-- proposing to configure memlock for mailbox process as well .

Before proposing to higher management for change in these settings to an optimized value, I wanted to confirm and take suggestions about this issue and few other parameters which could be adjusted.

Is disk IO running at 100 percent?

not exactly 100 % , but yes it goes as high as 70%

any one has any idea about it..?? please help..TIA.

Could you please post the output of the following commands, taken at a period with high load (like, when your backup runs). Ideally the following commands/scriptlets are started and stopped at the same time so they refer to the same period of time. Let them run for ~5 minutes:

Code:

vmstat 5

while : ; do ps -Ao vsz,pid,cmd | sort -rn | head 20 ; sleep 20 ; done

iostat -dxz 5

For iostat to be available you might have to install the "sysstat" rpm package.

bakunin

what is the your backup mechanisms ?

I think this is normal. 70-100% is not hight,while up 120% is another level.

SSSSSSSSSSSSSBBBBB

Moderator's Comments:

edit by bakunin: I don't know what your problem is, but you should solve it elsewhere, not here. This is just plain spam and neither do we like that here nor do we condone it. Consider yourself warned for the first and absolutely last time.