Hi,
I've just started seeing a really bizarre issue on my cluster (it appears to only affect the cluster). We have a perl script which makes a system call to cat to a file. This process used to take a couple on minutes now it takes about an hour...I've looked at all the normal tools, iostat, vmstat, nfsstat (since it is writing to a mounted area) and nothing seems out of place regarding resource utilization. The only strange thing I see is in top and the process.
527184 userid 42 0 2112K 196K WAIT 0:00 0.00% cat
It has an awful lot of WAIT...
=G=
Received on Wed Oct 22 2003 - 22:21:45 NZDT