Open Stack

Beyond what is and isn’t technically possible at the file system level there is always the problem that the user may have more data than can fit into the reduced disk.

I don’t want to take away useful functionality from folks if there are cases where it already works – mostly I just want to improve the user experience, and  to me the biggest problem here is the current failure mode where the user can’t tell if the request has been tried and failed, or just not happened at all for some other reason.

What if we introduced a new state of “Resize_failed” from which the only allowed operations are “resize_revert” and delete – so the user can at least get some feedback on the cases that can’t be supported ?

From: Aryeh Friedman [mailto:aryeh.friedman at gmail.com]
Sent: 13 June 2014 18:15
To: OpenStack Development Mailing List (not for usage questions)
Subject: Re: [openstack-dev] [nova] Do any hyperviors allow disk reduction as part of resize ?

Also ZFS needs to know what is on the guest for example bhyve (the only working hv for bsd currency [vbox kind of also works]) stores the backing store (unless bare metal) as single block file.   It is impossible to make that non-opaque to the outside world unless you can run commands on the instance.

On Fri, Jun 13, 2014 at 11:53 AM, Darren J Moffat <Darren.Moffat at oracle.com<mailto:Darren.Moffat at oracle.com>> wrote:


On 06/13/14 16:37, Daniel P. Berrange wrote:
The xenapi implementation only works on ext[234] filesystems. That rules
>out *BSD, Windows and Linux distributions that don't use ext[234]. RHEL7
>defaults to XFS for instance.
Presumably it'll have a hard time if the guest uses LVM for its image
or does luks encryption, or anything else that's more complex than just
a plain FS in a partition.

For example ZFS, which doesn't currently support device removal (except for mirror detach) or device size shrink (but does support device grow).  ZFS does support file system resize but file systems are "just" logical things within a storage pool (made up of 1 or more devices) so that has nothing to do with the block device size.

--
Darren J Moffat


_______________________________________________
OpenStack-dev mailing list
OpenStack-dev at lists.openstack.org<mailto:OpenStack-dev at lists.openstack.org>
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev



--
Aryeh M. Friedman, Lead Developer, http://www.PetiteCloud.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20140616/292190dc/attachment.html>