On Fri, 20 Apr 2012, Niels_Walet wrote:
When moving my virtual machines (libvirt/qemu-kvm) from one server to another (from amd to intel hardware), I seem to have suddenly hit the time-out issues that have been discussed in many places (the dreaded "blocked more than 120s" message), after which the systems both become totally unresponsive). Since the time-out involves the filesystem, I can only take screenshots, which I attach; nothing appears in the syslog.
This timeout issue is not specific to virtual machines, at Fermilab we see it on bare metal machines just as much as we do on virtual machines.
I have updated the virtual machines from SL 5.5 to 5.7; with some change but similar crashes; I added a few boot parameters (having to do with idle=), with no change at all. Does anyone have any suggestion what I could try to further diagnose this problem, or maybe even a solution? Niels Walet
More than half the time when I've seen those timeouts they have been blocking on some sort of network task or other. Do you have any kind of a network file system mounted such as NFS, AFS, GFS, etc? Steve Timm ------------------------------------------------------------------ Steven C. Timm, Ph.D (630) 840-8525 t...@fnal.gov http://home.fnal.gov/~timm/ Fermilab Computing Division, Scientific Computing Facilities, Grid Facilities Department, FermiGrid Services Group, Group Leader. Lead of FermiCloud project.