Re: [MirageOS-devel] Mirage tracing status

[Mindy <mindy@xxxxxxxxxxxxxxxxxxx>]

On 11/11/2014 10:53 AM, Thomas Leonard wrote:
> On 11 November 2014 16:40, Mindy <mindy@xxxxxxxxxxxxxxxxxxx> wrote:
> > On 11/11/2014 09:56 AM, Thomas Leonard wrote:
> > > [snip]
> > >
> > >
> > > If anyone has had any success using the tracing themselves (or got
> > > stuck), let me know!
> > I don't know that I'd call it "success" yet, but I've gotten some traces out
> > of unikernels; I'm now rewriting to use the interface you've just released.
> Cool :-)
>
> Note: to read the data from a Xen guest, you'll need to use the
> "collect" tool, currently in mirage-trace-viewer
> (https://github.com/talex5/mirage-trace-viewer), which can be built
> with:
>
> $ ./configure --enable-collect
>
> It's conditional because it depends on xenstore_transport.unix, which
> requires the Xen headers to be installed. I don't know if there's some
> existing tool for dumping a Xen shared memory region that I should be
> using instead.
Thanks very much for mentioning this. :)
> > The ability to read the trace buffer after crashes is very welcome :)
> Are you using AWS? Some extra work may be needed there, since you
> don't have access to dom0. Can you dump the RAM somehow using their
> API?
Luckily for me, what I'm trying to get at is reproducible locally, where 
I *do* have access to dom0.  I'd be surprised if this was possible in 
AWS EC2 (a very cursory glance suggests that it isn't).
-Mindy

_______________________________________________
MirageOS-devel mailing list
MirageOS-devel@xxxxxxxxxxxxxxxxxxxx
http://lists.xenproject.org/cgi-bin/mailman/listinfo/mirageos-devel