On Wed, 2011-08-10 at 17:08 +0400, Константин Алексеев wrote: > http://docs.ganeti.org/ganeti/2.2/html/install.html > "For optimum performance when running both CPU and I/O intensive > instances, it’s also recommended that the dom0 is restricted to one > CPU only, for example by booting with the kernel parameter nosmp." > > Also this benchmarking advises disable smp > http://web2.clarkson.edu/class/cs644/xen/files/repeatedxen-usenix04.pdf
Ignoring whether or not this is good advice (I expect it very much depends on your workload) this can be better achieved by adding dom0_max_vcpus=1 to your hypervisor command line or by hot-unplugging vCPUS once the system has booted (by echoing 0 into /sys/devices/system/cpu/cpu*/online for those CPUs). (the later give you the option of adding back VCPUs dynamically if you find you need to) That said, nosmp really ought to work I suppose so if this still happens on a recent kernel (e.g. 3.0) then I would appreciate it if you would also report this issue upstream (see http://wiki.xen.org/xenwiki/ReportingBugs) Did you confirm that this doesn't also happen under native? On first glance it looks more like an issue with a specific driver than a Xen issue. Ian. > > 2011/8/10 Ben Hutchings <b...@decadent.org.uk> > On Wed, 2011-08-10 at 16:51 +0400, Konstantin Alekseev wrote: > > Package: xen-linux-system-2.6.32-5-xen-amd64 > > Version: 2.6.32-35 > > Severity: normal > > > > > > With kernel options 'nosmp' dom0 hang > [...] > > Why does this matter to you? > > Ben. > > > > > -- > Konstantin -- Ian Campbell mixed emotions: Watching a bus-load of lawyers plunge off a cliff. With five empty seats. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org