Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I have a VMWare VM on an ESX host that has LVM partitions. I've configured kdump with very basic configuration - ext /dev/mapper/logical-volume-name and path /data/crash. When I force a system crash, it loads the kdump kernel and it shows the logical volumes, loads them and shows that it is saving the memory dump, and takes about 30 seconds to export about 2.5GB, which is normal. The problem is, when it reboots, it comes back up, but no network connectivity (can't get out, and can't get in via network), however, all network services are running and a simple reboot (no crash) fixes this. Additionally, another issue - the vmcore memory dump saved was not actually saved. I've tested this on a VM on VirtualBox and this works running a CentOS equivalant of the RHEL server that is currently in production that is currently having these kdump issues.

Any ideas or things I should look in to?

share|improve this question

1 Answer 1

I spoke to one of the developers for KDump and confirmed that this is a known bug. At this time, I cannot make any changes to production, but have a very suspicious idea that either the kernel needs to be upgraded and/or something between the VM and the VM Host need looked into.

from    Eric W. Biederman 
to  Drew Decker 
cc  jbarnes@sgi.com,
khalid.aziz@hp.com,
hari@in.ibm.com,
tjd21@cl.cam.ac.uk
date    Tue, Jan 25, 2011 at 1:50 PM
subject Re: KDump Question/Issues (Need a developer's eyes)

That does sound like a bug in the RHEL kernel interacting with VMware.

The fact that the reboot doesn't reset the hardware is frankly scary.

At this point any RHEL kernel is old and has not been updated by the
many many bug fixes that have happened between when they branched off
linus's tree and the present, and presumably comes with a support
contract by people that are paid to help you, and understand those
details, so I recommend you talk to the fine people at Red Hat.  They
should be able to help.

Usually problems like this come down to driver issues of some sort, not
issues in the actual kexec on panic infrastructure.  Although in this
case I would also suspect how the kdump kernel is calling reboot.

Good luck to you,
Eric
share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.