From: Eric Blake <eblake@redhat.com>
To: Kevin Wolf <kwolf@redhat.com>
Cc: Peter Maydell <peter.maydell@linaro.org>, qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [PATCH] Documentation: Warn against qemu-img on active image
Date: Thu, 16 Aug 2012 07:49:18 -0600 [thread overview]
Message-ID: <502CFA5E.50900@redhat.com> (raw)
In-Reply-To: <502CF711.2060000@redhat.com>
[-- Attachment #1: Type: text/plain, Size: 2221 bytes --]
On 08/16/2012 07:35 AM, Kevin Wolf wrote:
> Am 16.08.2012 14:56, schrieb Eric Blake:
>> On 08/16/2012 04:00 AM, Peter Maydell wrote:
>>> On 16 August 2012 10:00, Kevin Wolf <kwolf@redhat.com> wrote:
>>>> People have repeatedly expected that you can do things like snapshotting
>>>> an image with qemu-img while a qemu instance is running. Maybe we need
>>>> to consider locking the files while they are in use,
>>
>> Sounds like a nice feature bit to add to qcow2v3, where both qemu-img
>> and qemu check if the locking feature is enabled for an image, as well
>> as maintain a header bit that is set when the image is open read-write
>> and refuse to use the image if the lock bit is set.
>
> I thought the same. However, then you need some way to override this
> mechanism when recovering from a crash etc., so it's not a trivial
> addition and it would be user-visible.
Good point. With fcntl() locking, the lock goes away on crash; but with
file modification locking, the lock can get stuck so you have to provide
overrides; and once you provide overrides, the lock is not quite as
powerful. So maybe it's best to just leave locking up to management
apps (after all, libvirt already has a lock protocol support, currently
built on sanlock but also with a patch proposed for using fcntl()).
>> Is this strong enough? Remember, with qcow2v3 and qed, the mere act of
>> opening an image will perform refcount checks that modify the image,
>> unless you explicitly request otherwise, which means even a query of the
>> file metadata may result in modifying the image as part of the default
>> open.
>
> Not for read-only opens. I think qemu-img gets this right meanwhile, so
> that images are opened read-only when they are only queried.
Then it might also be worth documenting which actions are read-only
queries vs. potential modifications, in the docs for each action. (For
example, 'info' is read-only, 'convert' is read-only when creating a
copy although consistency is essential for it to be useful, 'check' is
readonly while 'check -r' is read-write,...)
--
Eric Blake eblake@redhat.com +1-919-301-3266
Libvirt virtualization library http://libvirt.org
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 620 bytes --]
prev parent reply other threads:[~2012-08-16 13:49 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-08-16 9:00 [Qemu-devel] [PATCH] Documentation: Warn against qemu-img on active image Kevin Wolf
2012-08-16 10:00 ` Peter Maydell
2012-08-16 12:56 ` Eric Blake
2012-08-16 13:35 ` Kevin Wolf
2012-08-16 13:49 ` Eric Blake [this message]
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=502CFA5E.50900@redhat.com \
--to=eblake@redhat.com \
--cc=kwolf@redhat.com \
--cc=peter.maydell@linaro.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.