Configure XEN Dom0 for a kernel dump

This document (7001383) is provided subject to the disclaimer at the end of this document.

Environment

SUSE Linux Enterprise Server 11 Service Pack 3 (SLES 11 SP3)
SUSE Linux Enterprise Server 10 Service Pack 2 (SLES 10 SP2)

Situation

Dom0 on a SLES 10 XEN server is crashing/panicking.

Resolution

Update the server to SLES 10 SP2 or newer and corresponding XEN packages (required). Always review the update notes as it may be necessary to update any SLES Virtual Machines (and Virtual Machine Driver Pack installs on any VMs) prior to the XEN Host (dom0) update; Especially if updating the XEN Host from SLES 10 (pre SP1).
Configure dom0 for kdump

Install the Yast2-kdump package on the XEN Host (dom0)
Configure dom0 for a kernel dump

From Yast2 > System > Kernel Kdump

Enable Kdump (keep defaults)

From Yast2 > System > Boot Loader

Highlight the XEN kernel, then click Edit
Cut and paste the crashkernel bootloader argument from the "Optional Kernel Command Line Parameter" section to the "Additional Xen Hypervisor Parameters" section
Note that for Xen on SLES 11, the syntax is still crashkernel=size@offset - e.g. crashkernel=512M@16M
So you may need to modify the argument from the native kernel when appending it to the xen.gz options.

Note: The makedumpfile tool currently only supports the "ELF" format and a dumplevel setting of "0". Check /etc/sysconfig/kdump file and verify KDUMP_DUMPLEVEL="0" and KDUMP_DUMPFORMAT="ELF". As of 2014 this is still the case for the current SLE versions. SUSE engineering has developed fixes to the makedumpfile and kdump packages for SLE 11 SP3, which allow setting the dumplevel to "1" to at least allow some level of page filtering to take place. These patches are expected to get released with our maintenance stream. Should a customer have an urgent need for this functionality in the interim period, please open a service request with SUSE support and refer to bug 864910.
Test by triggering a kernel dump

Shutdown all VMs and reboot the VM Server
From a terminal on dom0 run echo c > /proc/sysrq-trigger

Note: The system will become unresponsive until the core is written. When focus returns, verify that a vmcore has been written to the /var/log/dump directory.

Additional Information

See TID 3374462: Configure kernel core dump capture for more kdump info
See Troubleshooting a privileged (Dom0) Xen kernel for information on pre-SP2 XEN kernels.

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.