Hi, I'm experiencing similar problem on RHEL6. It's seems to be due to excessive context switches. The same VMs running on RHEL5 (KVM) didn't exhibit this problem. I found this issue described in the following bugzilla case:
https://bugzilla.redhat.com/show_bug.cgi?id=645898 When your VMs are running, check: vmstat 1 This supposed to be fixed in RHEL6.1 kernels. Regards, Greg Wojcieszczuk On 05/17/2011 03:30 PM, Dmitry Makovey wrote: > On Thursday, May 12, 2011, Dmitry Makovey wrote: >> Hi everybody, >> >> we have some weird problem with EL6 box which serves as a KVM host for 3 >> VMs currently. It shares the same interface with all 3 VMs (same network). >> All works, except from time to time machine behaves very slow in console >> (ok, SSH connection) and it happens only from time to time. >> >> We've got vnstat running on that box. There is no sign of network overload >> so far. Average load never crosses 1 (actually it's in 0.2 neighbourhood) >> so we're really at loss. >> >> Did anybody experience something similar? Any suggestions on where to look? > looking at responses is it safe to assume - nobody ever seen problems > described above (which would mean we have to dig into our configuration to be > able to spot the problem) > > > > -- > redhat-sysadmin-list mailing list > [email protected] > https://www.redhat.com/mailman/listinfo/redhat-sysadmin-list
-- redhat-sysadmin-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/redhat-sysadmin-list
