Xen project Mailing List

[Xen-users] Save/restore broken in 3.0.2-2?

From: "Simon Graham" <simgrxp@xxxxxxxxx>

Date: Mon, 8 May 2006 10:39:15 -0400

Delivery-date: Mon, 08 May 2006 08:50:06 -0700

Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:to:subject:mime-version:content-type; b=ehZSYvlL3TJHytKKjhHWr7u7Cn9HpZgk93Flk6bTI+Nz2WzNTjlRjqXexU9ZddZJxDUvTJlGF32pfPioMROqWYGllc6ZRLF6ZHovEt3jhdbMUHIGw+ZAJ2YgtGWawjEV/kxyvd/E5UIFasjZ+YU/XLjV4lhlCK9Y2Q0JJe3uFyQ=

List-id: Xen user discussion <xen-users.lists.xensource.com>

I've been experimenting with using save/restore as a poor-mans migration on 3.0.2-2 and have experienced some real problems with the guests hanging when they are restored. In the end, I was able to repro this simply executing 'xm save' and then 'xm restore' on the same system -- the symptom is that the the restored guest is apparently running (for example, I can ping it) but anything that requires access to the disk hangs up.

Eventually, I dumped the guest and used crash to look at what was going on and I can see that the VBD front end driver is hung waiting for disk I/O to complete -- it seems to reconnect to the backend OK and re-issues I/O that was outstanding when the guest was saved (in all cases, there seems to be a single I/O outstanding when the save is done - a syslog write) and even carries on and issues more disk I/O requests _but_ none of them complete.

I should say that both Dom0 and DomU are 3.0.2-2 on top of vanilla 2.6.16 kernels.

Anyone know if this is a known issue with the release? Or suggest further investigation required?

Simon

_______________________________________________ Xen-users mailing list Xen-users@xxxxxxxxxxxxxxxxxxx http://lists.xensource.com/xen-users