Corrupted double-linked list KVM error

[theNADS]

I am having an issue with a kvm guest (sl6.1) on a sl6.1 KVM host.

The install seemed to go ok. At the end of the install I clicked 'reboot' on the vnc installer and watched the virsh console for the guest whilst it rebooted.

The virtual machine did not reboot. I got this error:

Requesting system reboot
Restarting system

*** glibc detected *** /usr/bin/virsh: corrupted double-linked list: 0x000000000101b1c0 ***

To try and get the vm running I did a virsh destroy and then a virsh start of the vm - and the domain started!

The vm seems to run ok since I got it to boot, but the above error message looks as though it could be serious. Has anyone encountered this error? Is it a re-install of the vm?

sl5.6 installed fine.

jk

[helikaon]

Hi,
this looks like you hit a bug. Any progress on this?

cheers,

[theNADS]

I reinstalled the vm as i did not like to look of the error message.

Everything reinstalled ok using the same partitions (i just formated each partition and reinstalled the system in an identical way to the first install).

This time the machine rebooted without any error messages or issues.

However, I now find that 'virsh shutdown vm_guest_name' does not work!

Anyone had this issue before?

jk

[theNADS]

The virsh shutdown issue was nothing really. The guest vm just needed:

# yum install acpid
# chkconfig acpid on
# service acpid start

I think I am now finally up and running with KVM and ready to start configuring some guests.

[theNADS]

Just when I thought i was getting somewhere I have discovered that i still have some KVM issues.

When i first installed the sl6.1 guest on the sl6.1 host and noticed the glibc error I had a sl5.6 guest already running. I now believe it was the sl5.6 guest that is causing the errors that a am seeing; not the sl6.1 guest as i had first thought!

The messages that I see I /var/log/messages are shown below:

Sep 11 20:00:34 kvm-sl6x kernel: kvm: 2022: cpu0 unimplemented perfctr wrmsr: 0xc0010005 data 0x0
Sep 11 20:00:34 kvm-sl6x kernel: kvm: 2022: cpu0 unimplemented perfctr wrmsr: 0xc0010001 data 0x130076
Sep 11 20:00:34 kvm-sl6x kernel: kvm: 2022: cpu0 unimplemented perfctr wrmsr: 0xc0010005 data 0xffffffffffde6d26
Sep 11 20:00:34 kvm-sl6x kernel: kvm: 2022: cpu0 unimplemented perfctr wrmsr: 0xc0010001 data 0x530076
[root@kvm-sl6x ~]# cat /var/log/messages | grep rdm
Sep 11 14:05:14 kvm-sl6x kernel: kvm: 1928: cpu0 unhandled rdmsr: 0xc0010001
Sep 11 14:17:05 kvm-sl6x kernel: kvm: 1904: cpu0 unhandled rdmsr: 0xc0010001
Sep 11 19:06:33 kvm-sl6x kernel: kvm: 1896: cpu0 unhandled rdmsr: 0xc0010001

These messages also appear on the KVM host console. A sample of some of the additional error messages I am seeing in /var/log/messages are as follows:

Sep 10 21:34:06 kvm-sl6x libvirtd: 21:34:06.247: 9054: error : qemuSecurityDACRestoreSecurityFileLabel:84 : cannot resolve symlink /var/lib/libvirt/boot/virtinst-vmlinuz.sjq9W7: No such file or directory
Sep 10 21:34:06 kvm-sl6x libvirtd: 21:34:06.252: 9054: error : qemuSecurityDACRestoreSecurityFileLabel:84 : cannot resolve symlink /var/lib/libvirt/boot/virtinst-initrd.img.lUg3i0: No such file or directory
Sep 11 01:14:25 kvm-sl6x libvirtd: 01:14:25.067: 1766: error : virFDStreamRead:301 : cannot read from stream: Input/output error
Sep 11 01:14:25 kvm-sl6x libvirtd: 01:14:25.067: 1766: error : virFDStreamRead:301 : cannot read from stream: Input/output error
Sep 11 01:14:25 kvm-sl6x libvirtd: 01:14:25.068: 1766: error : virFDStreamRead:301 : cannot read from stream: Input/output error
Sep 11 01:18:27 kvm-sl6x libvirtd: 01:18:27.337: 1780: error : virGetHostname:2610 : internal error getaddrinfo failed for 'kvm-sl6x': Name or service not known
Sep 11 14:10:22 kvm-sl6x libvirtd: 14:10:22.383: 1770: error : qemudShutdownVMDaemon:3517 : Failed to send SIGTERM to vm_data (1953): No such process

I have tried an sl5.6 x86_64 guest, an sl5.6 i386 guest and an sl5.7 x86_64 guest. They all give me similar error messages on the host machines console.

The 32 bit sl5.6 guest also managed to crash the libvirt console connection resulting the the following error being displayed (once again):

*** glibc detected *** /usr/bin/virsh: corrupted double-linked list: 0x000000000101b1c0 ***

I cannot see any error messages in:

vi /var/log/libvirt/qemu/vm_data.log

Has this stuff been seen before? Is there any resolution, or should I be enquiring about this on the mailing list for Scientific Linux?

Google has let me down a little on this issue I am afraid!

jk

[theNADS]

I have done some more testing and the above rdmsr errors are actually related to restarting the sl6.1 guest!

[root@kvm-sl6x ~]# cat /var/log/messages | grep rdmsr
Sep 11 23:52:25 kvm-sl6x kernel: kvm: 1912: cpu0 unhandled rdmsr: 0xc0010001
Sep 11 23:54:53 kvm-sl6x kernel: kvm: 1994: cpu0 unhandled rdmsr: 0xc0010001
Sep 12 00:09:34 kvm-sl6x kernel: kvm: 2012: cpu0 unhandled rdmsr: 0xc0010001

An Ubuntu 10.04 guest has also now been installed on the sl6.1 kvm host and stopping and starting this guest gives no errors at all.

Is this a KVM bug or have I done something wrong here?

jk