From: pareilly@tcd.ie
Date: Tue Nov 25 2003 - 07:08:01 EST
I've got a problem with our Tru64 cluster mail server this morning, in
that many processes are stuck in a WAIT state. ps aux shows they are
in state U - however there's 700 MB RAM free, no swap is being used, CPU is
at 40%, so it appears to be disk. Even doing an 'ls -la' on a large directory
tree is very chunky in response. ie: display some output, wait 5 seconds,
display some more, wait 10 seconds etc.
iostat output shows
# iostat dsk10 dsk11 1
tty dsk10 dsk11 cpu
tin tout bps tps bps tps us ni sy id
1 250 5960 398 6996 474 19 0 30 51
0 52 6242 663 9143 784 13 0 62 25
0 52 7089 260 7728 333 18 0 37 45
0 52 10912 638 8320 603 20 0 46 34
0 51 3134 238 4496 269 10 0 68 22
0 52 9440 717 9688 697 23 0 66 11
0 50 3627 254 4880 325 15 0 77 7
This is on a HSG80 SAN, but nothing has changed so I'm not sure
how to diagnose what the processes are waiting on. Any advice on things
to chekc would be appeciated.
Paul
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:49:44 EDT