The xend logs from destination node are not very informative. They just tell
us the migrate was aborted by the source node. The one source xend log in
that bugzilla ticket indicates that shadow mode could not be enabled on the
source VM (for dirty page tracking). Could be a lack of memory if you are
using xend''s auto-ballooning feature (ie. dom0 boots with all memory;
gives
up memory as other VMs start). Suggestions: Look in your source machine xend
log (/var/log/xen/xend.log); try doing a non-live migrate; try disabling
xend auto-ballooning and just give dom0 the actual amount of memory it needs
to do its job up front (this is more logical if your dom0 is just a service
domain for other VMs; less logical if it is your main OS which happens to
run a few VMs on the side -- the latter case is likely to have more variable
memory demands).
-- Keir
On 1/9/08 02:58, "Tom Lanyon" <tom@netspot.com.au> wrote:
> Hi list,
>
> The bug reporting guide in the wiki mentions we should notify the -
> devel list when reporting a new bug, so I just wanted to bring some
> attention to:
>
> http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1328
>
> I''d reported another bug myself
> (http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1332
> ) and marked it as a duplicate of the one above as they seem to be the
> same issue, but can''t tell whether this is actually caused by the
> number of VCPUs or disk devices or something else entirely.
>
> Thanks and regards,
> Tom
>
> --
> Tom Lanyon
> Systems Administrator
> NetSpot Pty Ltd
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel