Most poeple get confused when it comes to configuring kdump for dom0. Since there are two entries in grub.conf, one representing the hypervisor and the other representing the dom0 kernel, the confusion is while thinking where should one give the "crashkernel" parameter. Have seen some people who are not ready to do trial and error passing the "crashkernel" parameter to both hypervisor and kernel. If you are too lazy to do a trial and error, this blog may help you.
The crashkernel parameter need to be passed to the hypervisor.
Eg, (Taken from my test system)
title Red Hat Enterprise Linux Server (2.6.18-92.1.18.el5xen)
root (hd0,0)
kernel /xen.gz-2.6.18-92.1.18.el5 crashkernel=128M@16M
module /vmlinuz-2.6.18-92.1.18.el5xen ro root=/dev/VolGroup00/LogVol00 rhgb quiet
module /initrd-2.6.18-92.1.18.el5xen.img
- Configure /etc/kdump.conf appropriately. The file is self descriptive.
- Run "service kdump propagate"
- Run "service kdump start"
3 comments:
Question 1
What kernels should be installed on server?
Question 2
What
memory for crash kernel (0x0 to 0x0) notwithin permissible range
1- For the kdump to work properly, both xen and non-xen kernel should be installed on the system.
2- I think you can safely ignore this message.
If you have configured Kdump, and expecting the system(kdump) to dump core when there is a kernel panic, you cannot safely ignore this warning."memory for crash kernel (0x0 to 0x0) notwithin permissible"
If this comes it won't dump core and you can't debug(if you intend to debug it.i.e)
Post a Comment