var sidebar_align = 'right';
var content_container_margin = parseInt('350px');
var sidebar_width = parseInt('330px');
WHM's suspension cause high load
We have a Dual opteron 275+4G ram box.
Normally, load is 2 as long as we do not perform any action on WHM
However, whenever we create/suspend/terminate an account , load burst to 25 and then httpd goes down
Admins say it is hardware issues, which does not sound right to me
have you guys got any clues on this wield issue?
I've never seen that happen. You would have to post a copy of TOP output. Sounds to me like it might be iowait skyrocketing, and would be a hardware problem. What is the RAID setup (if any)
here you go
top - 09:58:31 up 10 days, 12:46, 1 user, load average: 2.23, 2.78, 3.24
Tasks: 173 total, 1 running, 171 sleeping, 0 stopped, 1 zombie
Cpu(s): 3.2%us, 1.9%sy, 9.9%ni, 80.8%id, 4.0%wa, 0.0%hi, 0.2%si, 0.0%st
Mem: 4141628k total, 3476140k used, 665488k free, 116332k buffers
Swap: 2096472k total, 88k used, 2096384k free, 2452932k cached
I mean when the load is at 25
Check this item - '4.0%wa'
alright, will check it when we need to do anything
My suspicion is a huge bottleneck in writing to the hard drive(s). Even a load of 2-3 on the server which is basicly doing nothing is suspicious that that is the case. Is it a RAID5?
nope, there is no raid on server
It is SATA HD, should not be so weak though...
top - 10:31:39 up 10 days, 13:19, 1 user, load average: 2.27, 2.19, 2.36
Tasks: 221 total, 5 running, 216 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.0%us, 24.4%sy, 3.5%ni, 20.9%id, 51.0%wa, 0.0%hi, 0.2%si, 0.0%st
gee...high to 50%
Maybe try this and check the output:
hdparm -tT /dev/sda
It will do a basic hard drive test
Timing cached reads: 2972 MB in 2.00 seconds = 1486.97 MB/sec
Timing buffered disk reads: 290 MB in 3.01 seconds = 96.30 MB/sec
Timing cached reads: 3848 MB in 2.00 seconds = 1925.76 MB/sec
Timing buffered disk reads: 70 MB in 3.10 seconds = 22.62 MB/sec
That's not too bad. But I've never seen that happen just by terminating an account.
Maybe check the message log:
tail -f /var/log/messages
If you see a lot of hardware errors there's something wrong.
Or rather than sit watching the log you could do this:
grep error /var/log/messages
em...provider suggested me to increase query cache of mysql, Iwill see how it goes
tested increasing ram, no works....
I found that IO wait burst when proeeding
do we have to run this everytime suspending? if not, how to disable it?
Maybe run /scripts/fixquotas but yes, the closer you keep WHM to stock, the easier. You can turn quotas off in WHM (I think) but no one will have a disc space limit.
This post might help:
But my servers all run fine with quotas being on /var partition.
Last edited by page-zone; 03-17-2008 at 11:22 PM.
this is a great info
thanks so much for this, i will let admins look into it