Martin Troester
2010-Apr-13 20:44 UTC
[Xen-users] HVM DomU with Kernel 2.6.27-19 on CentOS 5 hangs with 100% CPU at Linux bootup on Xen 3.2.1
Dear Xen Community, today we tried to upgrade our base server from a Xeon E5430 to a Xeon X5570 (Nehalem). This worked fine so far, but our DomUs with Kernel 2.6.27 no longer boot - they start up until the following message appears: [..] Mount-cache hash table entries: 256 Initializing cgroup subsys ns Initializing cgroup subsys cpuacct Initializing cgroup subsys memory Initializing cgroup subsys devices Initializing cgroup subsys freezer Afterwards, the CPU utilization in xm top/xm list shows one VCPU assigned to that VM, even if more have been specified in the config file. This VCPU is 100% busy - so it appears to be caught in some kind of a loop? As only one VCPU is shown at this point, it appears to be early during bootup, not all VCPUs have been detected and assigned to that VM. As the very same image runs without problems with the same Dom0 and DomU config on the "old" Xeon, and other DomUs still run on the new setup, this seems to be related to this kernel version. Any help to debug deeper into this are highly appreciated. Thanks! Martin _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Martin Troester
2010-Apr-13 20:53 UTC
[Xen-users] HVM DomU with Kernel 2.6.27-19 on CentOS 5 hangs with 100% CPU at Linux bootup on Xen 3.2.1
Dear Xen Community, today we tried to upgrade our base server from a Xeon E5430 to a Xeon X5570 (Nehalem). This worked fine so far, but our DomUs with Kernel 2.6.27 no longer boot - they start up until the following message appears: [..] Mount-cache hash table entries: 256 Initializing cgroup subsys ns Initializing cgroup subsys cpuacct Initializing cgroup subsys memory Initializing cgroup subsys devices Initializing cgroup subsys freezer Afterwards, the CPU utilization in xm top/xm list shows one VCPU assigned to that VM, even if more have been specified in the config file. This VCPU is 100% busy - so it appears to be caught in some kind of a loop? As only one VCPU is shown at this point, it appears to be early during bootup, not all VCPUs have been detected and assigned to that VM. As the very same image runs without problems with the same Dom0 and DomU config on the "old" Xeon, and other DomUs still run on the new setup, this seems to be related to this kernel version. Any help to debug deeper into this are highly appreciated. Thanks! Martin _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Martin Troester
2010-Apr-18 20:48 UTC
Re: [Xen-users] HVM DomU with Kernel 2.6.27-19 on CentOS 5 hangs with 100% CPU at Linux bootup on Xen 3.2.1
Hi, sorry for the double-posting. The solution for us to get over this problem was to upgrade to Xen 3.4.2. Before doing the upgrade of the hypervisor, I verified that upgrading the DomU kernel to 2.6.32 did not help either. Hope this helps anyone who sees the same issue. Regards, Martin _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users