have two virtual servers (guests and host ubuntu 606) and I'm seeing excessively high load averages in both of them. vmware server 1.0.2. System is a moderately old dual Intel CPU system. The OS disc is IDE but the virtual machine discs reside on a SCSI connected raid five box. The load average problem seems to be tied into some aspect of disk I/O activity. Here are some sample alerts I receive via the hobbit monitoring tool. apologies for the crappy formatting. Apparently the editor for this web forum written by people that never used preformatted text.
Notice that the CPU utilization is very low but the load average is 35 in the first one and 51 in the second period if you sit there watching top during one of these events, you see that the idle time goes to zero and the wait time spikes to 70 or 80%. The host load average climbs to roughly one half of the guest load average and shows the same decrease in idle time and and spikes in wait time.
I have seen the same behavior with vmware server 1.0.1-1.0.4
Any ideas as to what is happening and how do I fix it?
-
red Mon Nov 19 07:23:48 EST 2007 up: 205 days, 0 users, 170 procs, load=17.25
&red Load is CRITICAL
System clock is -545 seconds off
top - 07:24:52 up 205 days, 21:07, 0 users, load average: 35.69, 17.25, 6.63
Tasks: 170 total, 1 running, 169 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.5% us, 4.3% sy, 0.0% ni, 90.8% id, 3.4% wa, 0.2% hi, 0.8% si
Mem: 1572108k total, 1457776k used, 114332k free, 353812k buffers
Swap: 377488k total, 120k used, 377368k free, 870868k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
17706 hobbit 15 0 2192 1040 752 R 3.5 0.1 0:00.11 top
25237 www-data 25 0 223m 2624 1228 S 1.7 0.2 0:00.18 apache2
17666 dgreco 15 0 2556 1376 1060 S 1.7 0.1 0:00.19 imap
1 root 15 0 1468 472 412 S 0.0 0.0 1:35.78 init
2 root RT 0 0 0 0 S 0.0 0.0 0:00.00 migration/0
-
red Mon Nov 19 11:37:57 EST 2007 up: 206 days, 0 users, 255 procs, load=29.40
&red Load is CRITICAL
System clock is 341 seconds off
top - 11:41:30 up 206 days, 1:19, 0 users, load average: 51.24, 29.31, 12.43
Tasks: 251 total, 1 running, 249 sleeping, 0 stopped, 1 zombie
Cpu(s): 0.5% us, 4.3% sy, 0.0% ni, 90.8% id, 3.4% wa, 0.2% hi, 0.8% si
Mem: 1572108k total, 1403264k used, 168844k free, 349376k buffers
Swap: 377488k total, 120k used, 377368k free, 800232k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
12757 bcollins 15 0 2380 1032 868 S 9.5 0.1 0:00.06 imap
12755 dovecot 17 0 3220 1436 1196 S 7.9 0.1 0:00.05 imap-login
12756 root 17 0 0 0 0 Z 7.9 0.0 0:00.05 dovecot-a <defunct>
12754 hobbit 15 0 2328 1096 752 R 6.4 0.1 0:00.20 top