Guest shutdown problem with kexec --load ... ; kexec --exec running under Xen on CentOS 7 (x86_64)

Dave Young dyoung at redhat.com
Tue Jul 25 22:16:01 PDT 2017


Cc Vitaly Kuznetsov, who is the expert of kdump on xen guest issues, but
I'm not sure if he ever tried kexec reboot..
 
On 07/21/17 at 01:27pm, Corey Ashford wrote:
> The platform is a PVHVM guest of CentOS 7 (x86_64) running on a XenServer
> (version 6.5 and 7.2 both have the same behavior).
> 
> This is the kexec-tools rpm build we are using:
> 
> kexec-tools-2.0.7-50.el7.x86_64
> 
> If we run these two commands:
> 
> kexec --load /boot/vmlinux-<current kernel> --initrd=/boot/initramfs
> <corresponding initramfs for current kernel> --reuse-cmdline
> 
> kexec --exec
> 
> The first line succeeds without errors, but the second one causes the VM to
> immediately shutdown, and there's no further output on the console. We've
> tried adding a number of different kernel command line options, such as:
> debug=1 early_printk=xen loglevel=9
> 
> We've also tried blacklisting the PV drivers.
> 
> This sequence works fine on KVM, VMware, and also on bare-metal machines.
> 
> Is this expected behavior on Xen with CentOS 7.x?  We have scoured the
> kexec-tools mailing list, googled extensively, and haven't come up with much
> help.  We've also tried changing various kernel config parameters related to
> Xen, but the behavior doesn't change.
> 
> Any info, debugging advice, suggestions would be appreciated.
> 
> Thanks for your consideration,
> 
> - Corey
> 
> Corey Ashford
> IBM Embedded Linux, software engineer
> cjashfor at us.ibm.com
> 
> 
> _______________________________________________
> kexec mailing list
> kexec at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/kexec



More information about the kexec mailing list