Hello !
I have a very strange problem:
My GSX server has some weeks of uptime now and there are several VMs running. (Windows, Suse 9.1, Suse 9.3 ....)
i don`t know when it started and i`m not sure if this is from the very beginning when i moved the vm to this machine - but whenever a user generates some load inside the suse 9.1 VM, the whole system including host and all guests becomes extremly sluggish
This is no I/O issue because typing on the console is sluggish, too.
this seems to be a scheduling issue - but i don`t have a clue why this happens
when i start a cpu-hog (while true;do true; done) inside the suse 9.1 VM, top tells me the following:
Cpu0 : 0.3% us, 3.8% sy, 0.0% ni, 76.1% id, 19.7% wa, 0.0% hi, 0.0% si
Cpu1 : 0.0% us, 100.0% sy, 0.0% ni, 0.0% id, 0.0% wa, 0.0% hi, 0.0% si
we see, that %sys for cpu1 is at 100
when i do the same with any other system (i.e. suse 9.3 vm), the sluggishness doesn`t occur and all is fine - the load is equally distributed between both cpu`s
Cpu0 : 0.0% us, 61.7% sy, 0.0% ni, 38.0% id, 0.3% wa, 0.0% hi, 0.0% si
Cpu1 : 1.0% us, 42.9% sy, 0.0% ni, 56.1% id, 0.0% wa, 0.0% hi, 0.0% si
strange enoug - on another system with vmware-server i also see cpu`s at 100% sys, but don't recognize any sluggishness.
host is suse 9.3 (2.6.11.4-20a-bigsmp) - boot options are: root=/dev/sda2 vga=791 selinux=0 splash=normal resume=/dev/sda1 showopts elevator=cfq
problematic guest is suse 9.1 (2.6.4-52-default) - boot options are: "resume=/dev/sda1 showopts"
the configuration for the suse 9.1 and 9.3 VMs are nearly the same.
The "sluggishness" feels the same like when i once accidentally ran a VM which had a smp-kernel
whatever - the fact that a single VM kann kill the whole performance of my server that much seems like bug to me - but i canno`t tell if this is a vmware or a host kernel problem.
can somebody probably help with this ?
regards
roland