waitio: what process is causing this?

Hi,
sometimes my system redhat el as 3.4 goes in waitio 100% and i'm not finding what is doing that.
Is there a way to find the process causing so much io?

Thanks

I've heard of an AS 3 and AS 4, but not 3.4?

Unfortunately it's quite difficult to identify processes that are generating lots of I/O... it would be very handy if there was an iotop programme out there, but I don't think there is.  

Anyway, if a process is generating a lot of I/O, chances are that it is using a bit of CPU as well, so have a look in top and see if you can see any processes that are likely to be doing that.  And if, when they complete, the I/O wait goes away, then you have most likely identified the culprit.  

Annihilannic.

As Annihilannic mentioned, it can difficult to find the root cause.
Check for disk errors
Check your paging space usage (pi and po on vmstat)
Check which disk is high in I/O etc. etc.

NMON is an IBM freeware utility for AIX and Linux systems.
Great tool for understanding what's happening under the hood...

http://www-941.haw.ibm.com/collaboration/wiki/display/WikiPtype/nmon

"If you always do what you've always done, you will always be where you've always been."

Hi,
for redhat 3.4 I mean release 3 update 4.
I'm using top, vmstat, ps but not finding process using so much io.
On this system there's lotus domino 6.5.4 running, so I see multiple domino "server" processes running but no one is using more than 2% cpu.
System uses swap always nearly 55%.
cat /proc/swaps
Filename                        Type            Size    Used    Priority
/dev/sda8                       partition       4096472 684044  -1
/dev/sda9                       partition       4096472 0       -2
cat /proc/meminfo
        total:    used:    free:  shared: buffers:  cached:
Mem:  3982761984 3961831424 20930560        0 53850112 3674484736
Swap: 8389574656 701980672 7687593984
MemTotal:      3889416 kB
MemFree:         20440 kB

No errors in /var/log/messages.

Thanks again.

Have you tried using iostat -x to identify which filesystem/partition is under heavy I/O load?  Perhaps that will give you an idea which process is causing it.

Annihilannic.

Hi,
I know that the filesystems that have problems are the one related to notes (where reside db's data and archive).
What I want to know is what process is exactly blocking the system.
Sometimes waitio goes 100% and the only way to solve the problem is to restart the server!
I know it's something related with notes but don't know what is really blocking.

Thanks again.