[Pkg-xen-devel] Bug#912975: xen-hypervisor-4.8-amd64: Dom0 crashes randomly without logs on Debian Stretch with Xen 4.8.4

[Hans van Kranenburg]

Hi Alexander,

On 2/22/19 9:41 PM, Alexander Dahl wrote:
> 
> On Fri, Feb 22, 2019 at 07:24:11PM +0100, Hans van Kranenburg wrote:
>> The current state of this bug does not really allow anyone other than
>> yourself to cause it to progress.
> 
> FWIW, I also have problems with Xen and stretch on amd64. Since
> upgrading from jessie I get random crashes, which means the system
> hangs and I only can do a hard powercycle.

Ok. I'm gonna sound a bit strict/rigorous/stern here (I don't know which
of those is the right one, not a native speaker).

Please don't use an existing bug for an "I'm also having similar
problems" report. It might seem helpful to group similar symptoms
together, but often there seem to be different causes in the end, and
people from the package maintainer team will get confused about what the
real issue was and if it's fixed now or not, and you might end up with a
closed bug report while your issue was not dealt with, or the original
reporter might end up with a closed bug because your me-too problem was
fixed.

>From your problem description, it seems you're experiencing dom0 or even
xen hypervisor problems, not domU (virtual machine) problems. Is that right?

> I'm currently reorganizing
> the partitions to get enough space for a debug kernel on the rootfs,
> otherwise the stacktraces are probably not of big help?

As long as you don't share any stack trace at all, they won't be of any
help no. :) You might be experiencing a known problem, or a problem that
we know already has been fixed upstream in later Xen or Linux, or a new
problem.

> (I would have upgraded to buster already, but pygrub is broken there,
> so maybe we get stretch fixed until then.)

The pygrub fixes are in the upload to unstable that was done today.

https://tracker.debian.org/news/1031793/accepted-xen-411126-g87f51bf366-2-all-amd64-source-into-unstable/

Hans

Off-the-record: 2018 was not a good year for the Linux kernel in
general, also thanks to all the spectre/meltdown things happening. My
own experience is that the Linux 4.9 LTS kernel is quite unusable (as
dom0 and domU) with Xen, and I jumped over it, towards Xen 4.11 and
Linux 4.19, which is a great success so far.

So, with my personal (and $dayjob) hat on, I can recommend leaving the
current situation behind and at least please run the Linux 4.19 kernel
from stretch-backports instead.

With my Debian Xen package mainainer hat on... Yes, I'd like to help
you, but please create an additional bug after you have been able to
collect more logs and stacktraces and explosions happening etc.

Thanks.