Hello,
last weekend some production servers running Xen-3.4.3 and Linux-2.6.32
rebootet. Currently its unclear if the hardware is faulty or if its a
software problem. Is there an easy way to switch the Xen-hypervisor not to
reboot the the machine if the dom0 fails?
Doing a test with an "echo c >/proc/sysrq-trigger" crashes the
dom0, which
then makes the the hypervisor to reboot the machine. It would help if the
system would just stay in that faulty state and not reboot, so one could at
least look at the screen and see some kernel dump or similar.
I also tried to use kexec, but that doesn''t seem to work with Debian
kexec-tools, which complains about the _dom0_-Kernel not being prepared for
loading a crashdump kernel, even I pass "crashkernel=512M" via Grub to
the
_hypervisor_. (I really would prefer the Hypervirsor to switch to the
crashdump kernel instead of just the dom0-kernel, since I also would like to
get a hypervisor crashdump in addition to only a dom0 crashdump, since
I''m
unsure if the reboot is caused by a faulty dom0 or a Xen hypervisor problem).
Any help is appreciated.
Sincerely
Philipp
--
Philipp Hahn Open Source Software Engineer hahn@univention.de
Univention GmbH Linux for Your Business fon: +49 421 22 232- 0
Mary-Somerville-Str.1 D-28359 Bremen fax: +49 421 22 232-99
http://www.univention.de/
_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users
I also got into sitatsiyu when dom0 is half dead condition. And besides reset button - nothing could help me. What to do when dom0 freezes? How to override it with the commands if there is still an open server console but xend does not work anymore? PH> last weekend some production servers running Xen-3.4.3 and Linux-2.6.32 PH> rebootet. Currently its unclear if the hardware is faulty or if its a PH> software problem. Is there an easy way to switch the Xen-hypervisor not PH> to reboot the the machine if the dom0 fails? PH> Doing a test with an "echo c >/proc/sysrq-trigger" crashes the dom0, PH> which then makes the the hypervisor to reboot the machine. It would PH> help if the system would just stay in that faulty state and not reboot, PH> so one could at least look at the screen and see some kernel dump or PH> similar. PH> I also tried to use kexec, but that doesn''t seem to work with Debian PH> kexec-tools, which complains about the _dom0_-Kernel not being prepared PH> for loading a crashdump kernel, even I pass "crashkernel=512M" via Grub PH> to the _hypervisor_. (I really would prefer the Hypervirsor to switch PH> to the crashdump kernel instead of just the dom0-kernel, since I also PH> would like to get a hypervisor crashdump in addition to only a dom0 PH> crashdump, since I''m unsure if the reboot is caused by a faulty dom0 or PH> a Xen hypervisor problem). PH> Any help is appreciated. -- I try to select the top of the correspondence and collected in one location. Come and help comments http://lixen.ua-ohosting.com/
Hello, to answer my own questions: On Wednesday 25 January 2012 08:44:59 Philipp Hahn wrote:> last weekend some production servers running Xen-3.4.3 and Linux-2.6.32 > rebootet. Currently its unclear if the hardware is faulty or if its a > software problem. Is there an easy way to switch the Xen-hypervisor not to > reboot the the machine if the dom0 fails?Add "noreboot=1" to the Hypervisor options.> I also tried to use kexec, but that doesn''t seem to work with Debian > kexec-tools,xexec/kdump was only implemented in the non-PV-Ops Linux-2.6.18 series and never forward-ported to the newer PV-Ops-kernels. Sincerely Philipp -- Philipp Hahn Open Source Software Engineer hahn@univention.de Univention GmbH Linux for Your Business fon: +49 421 22 232- 0 Mary-Somerville-Str.1 D-28359 Bremen fax: +49 421 22 232-99 http://www.univention.de/ _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users