All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jamie Lokier <jamie@shareable.org>
To: qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [PATCH] qcow3 - arbitrary metadata
Date: Tue, 29 Jul 2008 07:44:12 +0100	[thread overview]
Message-ID: <20080729064412.GA32582@shareable.org> (raw)
In-Reply-To: <488E8AE8.2060602@codemonkey.ws>

Anthony Liguori wrote:
> >Yeah, I have the opposite problem - too much in the same file :-)
> >
> >I want to be able to savevm, but some of my VMs don't have any qcow2
> >images (because I don't trust them for mission-critical VMs since
> >recent discussion,
> 
> How do you snapshot the disk images?

Does savevm snapshot the disk images too?  I thought it was just the
machine state.  Does that mean non-qcow2 disks (e.g. raw files) are
not snapshotted but qcow2 ones are, in the same VM?

It's actually just the machine state I want to save.  I want to use
savevm to save the state of the VM before quitting, and loadvm to
restart.  Like suspend-to-RAM, but doesn't require the guest to
cooperate.  (Does suspend work anyway?)

> >That seems a bit hackish.  It would be nice if savevm could just write
> >to a named file and loadvm could read it.
> 
> I'm in the process of cleaning up the live migration code from KVM for 
> QEMU.  It works quite happily right now, I'm just reorganizing some things.

Excellent, KVM's migration makes a lot of sense in QEMU.
Thank you.

> In the process, I hope to introduce a checkpoint command which provides 
> a live snapshotting mechanism.  I think there is certainly room for 
> discussion about whether you should be able to snapshot/savevm to a file 
> besides snapshotting to a qcow2 file.

Another thing I wanted, but gave up on, was non-destructive work in a
VM, and if I like the work, commit it *without* pausing the VM for ages.

What typically happens is I use -snapshot or branch a temporary qcow2
off my main disk image.  That's all fine.

Then I do 'commit' in the monitor when I'm happy with my changes, and
the VM pauses for a long time while it writes back the changes.

Even just a few minutes work in Windows XP translates to a large
number of megabytes of changes to write back, it takes a while.

What I'd like is -snapshot or temporary qcow2 branch, and
'async_commit' which is my way of saying "I'm happy with the work done
so far, commit it, but I'd like to carry on working".

It would need to be atomic to be safe, I guess, and that begs for a
new qcow format.  I'm dabbling with database backends for another
project at the moment, who knows, watch this space.

-- Jamie

  reply	other threads:[~2008-07-29  6:44 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-07-28 19:56 [Qemu-devel] [PATCH] qcow3 - arbitrary metadata Nathaniel McCallum
2008-07-28 20:08 ` Anthony Liguori
2008-07-28 20:24   ` Nathaniel McCallum
2008-07-28 20:36     ` Paul Brook
2008-07-28 20:52       ` Anthony Liguori
2008-07-28 20:50     ` Anthony Liguori
2008-07-28 21:12       ` Nathaniel McCallum
2008-07-28 21:17         ` Anthony Liguori
2008-07-28 21:19           ` Nathaniel McCallum
2008-07-28 21:39             ` Anthony Liguori
2008-07-29  1:49   ` Jamie Lokier
2008-07-29  2:11     ` Jorge Lucángeli Obes
2008-07-29  2:48       ` Nathaniel McCallum
2008-07-29  3:05         ` Jorge Lucángeli Obes
2008-07-29  3:13     ` Anthony Liguori
2008-07-29  6:44       ` Jamie Lokier [this message]
2008-07-29 13:38         ` Anthony Liguori
2008-07-29 16:05           ` Jamie Lokier
2008-07-29  8:48     ` Daniel P. Berrange
2008-07-28 23:04 ` Laurent Vivier
2008-07-29  1:25   ` Nathaniel McCallum
2008-07-29  3:06     ` Anthony Liguori
2008-07-29  3:22       ` Nathaniel McCallum
2008-07-29  6:40         ` maht
2008-07-29  6:44           ` maht
2008-07-29  7:48     ` Laurent Vivier

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20080729064412.GA32582@shareable.org \
    --to=jamie@shareable.org \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.