kdump fails on Hyper-V
This document (7014585) is provided subject to the disclaimer at the end of this document.
Environment
SUSE Linux Enterprise Server 11 Service Pack 2 (SLES 11 SP2)
SUSE Linux Enterprise Server 11 Service Pack 3 (SLES 11 SP3)
Microsoft Hyper-V
SUSE Linux Enterprise Server 11 Service Pack 3 (SLES 11 SP3)
Microsoft Hyper-V
Situation
When running SUSE Linux Enterprise Server 11 SP2 or SP3 as a guest on Microsoft Hyper-V and a kernel core dump (kdump) is triggered, it fails to complete due to failure to find the root disk.
The errors received are similar to:
Could not find /dev/disk/by-id/scsi-14d5346542020202060bc93c5e99f6a4eb383477d7443364a-part2
Want me to fall back to /dev/disk/by-id/scsi-14d5346542020202060bc93c5e99f6a4eb383477d7443364a-part2? (Y/n)
The errors received are similar to:
Could not find /dev/disk/by-id/scsi-14d5346542020202060bc93c5e99f6a4eb383477d7443364a-part2
Want me to fall back to /dev/disk/by-id/scsi-14d5346542020202060bc93c5e99f6a4eb383477d7443364a-part2? (Y/n)
Resolution
As per Microsoft document http://support.microsoft.com/kb/2858695 , you must modify /etc/sysconfig/kdump to append ata_piix.prefer_ms_hyper_v=0 to KDUMP_COMMANDLINE_APPEND.
So, this line should read:
KDUMP_COMMANDLINE_APPEND="ata_piix.prefer_ms_hyper_v=0"
It is also necessary to change the partitions to mount by UUID instead of by Device ID.
This can be accomplished using the YaST Partitioner.
Open the YaST Partitioner by going to YaST->System->Partitioner.
A warning will be displayed indicating that partitioning disks that are in use may
result in data loss. It is safe to answer 'Yes'.
Select the partition, choose 'Edit', and select 'Fstab Options' under 'Mounting Options'
Under 'Mount in /etc/fstab by', change from Device ID to UUID, and select 'OK'.
Repeat this for every partition required to boot and any partitions required to store the
kernel core.
After this change, verify that both /etc/fstab and the grub boot loader configuration, /boot/grub/menu.lst refer to the root partition by UUID, run 'mkinitrd' and reboot.
So, this line should read:
KDUMP_COMMANDLINE_APPEND="ata_piix.prefer_ms_hyper_v=0"
It is also necessary to change the partitions to mount by UUID instead of by Device ID.
This can be accomplished using the YaST Partitioner.
Open the YaST Partitioner by going to YaST->System->Partitioner.
A warning will be displayed indicating that partitioning disks that are in use may
result in data loss. It is safe to answer 'Yes'.
Select the partition, choose 'Edit', and select 'Fstab Options' under 'Mounting Options'
Under 'Mount in /etc/fstab by', change from Device ID to UUID, and select 'OK'.
Repeat this for every partition required to boot and any partitions required to store the
kernel core.
After this change, verify that both /etc/fstab and the grub boot loader configuration, /boot/grub/menu.lst refer to the root partition by UUID, run 'mkinitrd' and reboot.
Cause
Different drivers are used in the primary kernel and the kdump kernel. This difference in drivers
cause the /dev/disk/by-id links to be different between the primary kernel and kdump kernel,
which causes the failure.
cause the /dev/disk/by-id links to be different between the primary kernel and kdump kernel,
which causes the failure.
Disclaimer
This Support Knowledgebase provides a valuable tool for SUSE customers and parties interested in our products and solutions to acquire information, ideas and learn from one another. Materials are provided for informational, personal or non-commercial use within your organization and are presented "AS IS" WITHOUT WARRANTY OF ANY KIND.
- Document ID:7014585
- Creation Date: 19-Feb-2014
- Modified Date:28-Sep-2022
-
- SUSE Linux Enterprise Server
For questions or concerns with the SUSE Knowledgebase please contact: tidfeedback[at]suse.com
