Xen Kernel cannot boot due to ATA erros/timeouts

Bug #1127214 reported by Anes Lihovac
12
This bug affects 2 people
Affects Status Importance Assigned to Milestone
xen (Ubuntu)
Confirmed
Undecided
Unassigned

Bug Description

In week 6 or 7 after an system update , my Xen host stopped booting.
When booting up the Xen Kernel, it will timeout and drop to initramfs shell. DMESG will show ATA errors/timeouts not finding my harddisk(s). The same system with the same Kernel but xen-less, boots and works without problems.

I tried several things :
1. adding all the modules which my xen-less system would have loaded to the initramfs -> no change
2. switching from ahci to ide mode in bios -> no success
3. changing uuid for root with its real block ids (like sda3, etc) in grub -> no change in boot behaviour
4. blkid in initramfs shell shows nothing !
5. boot Xen with an older installed Kernel Version (3.2.0-23-generic ) -> same problem
6. cat /proc/partitions shows nothing besides my cdrom(sr0)
7. recreate initrds
8. reinstall whole 12.04

I was able to save the DMESG output from the initramfs shell. The system would recognize a plugged in USB thumb drive, which I just needed to mount manually.

Due to the fact that it worked before updates fromn week 6 or 7 (soory I am not sure) but doesn't work when choosing an older kernel, it looks like it is not up to the Kernel, but something else, which I am affraid I cannot find.

Revision history for this message
Anes Lihovac (anes-lihovac-gmail) wrote :
Revision history for this message
Anes Lihovac (anes-lihovac-gmail) wrote :

I just compiled vanilla xen-4.2.1. This hypervisor version will boot as normal as it had it before week 6 or 7. I guess some update broke something in the Ubuntu hypervisor package.

Revision history for this message
Anes Lihovac (anes-lihovac-gmail) wrote :

After checking again 4.2.1, I am prett sure this is due to changes in 4.1.x xen-hypervisor in Ubuntu repos.

Revision history for this message
Launchpad Janitor (janitor) wrote :

Status changed to 'Confirmed' because the bug affects multiple users.

Changed in xen (Ubuntu):
status: New → Confirmed
Revision history for this message
Simon Déziel (sdeziel) wrote :

My Saucy dom0 is also affected by this. I wasn't able to get the dmesg in text but I attached a picture of it.

The problem seems to be this:

 ata1.00: qc timeout (cmd 0xec)
 ata1.00: failed to IDENTIFY (I/O error, err_mask=0x4)

Then the SATA link back down a on speed (6 -> 3 Gbps) and the same occurs over and over.

The same kernel boots just fine when outside of Xen.

System info:

$ lsb_release -rd
Description: Ubuntu 13.10
Release: 13.10

$ apt-cache policy linux-image-`uname -r` xen-hypervisor-4.3-amd64
linux-image-3.11.0-13-generic:
  Installed: 3.11.0-13.20
  Candidate: 3.11.0-13.20
  Version table:
 *** 3.11.0-13.20 0
        500 http://archive.ubuntu.com/ubuntu/ saucy-updates/main amd64 Packages
        500 http://security.ubuntu.com/ubuntu/ saucy-security/main amd64 Packages
        100 /var/lib/dpkg/status
xen-hypervisor-4.3-amd64:
  Installed: 4.3.0-1ubuntu1
  Candidate: 4.3.0-1ubuntu1
  Version table:
 *** 4.3.0-1ubuntu1 0
        500 http://archive.ubuntu.com/ubuntu/ saucy/universe amd64 Packages
        100 /var/lib/dpkg/status

Revision history for this message
Simon Déziel (sdeziel) wrote :

Here is the dmesg of the *non dom0* kernel that boots properly.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.