> > > > Or can we simply add a header to each write with additional information?
> > >
> > > Let's discuss it once we have a clearer picture of what that
> > > additional info would be.
> >
> > OK, just discovered such case.
> >
> > For example it would be nice to pass the size of the devices someh
On Mon, Feb 25, 2013 at 04:35:09PM +, Dietmar Maurer wrote:
> > >>> If we use nbd, how can we pass additional information to the other
> > >>> side, for example information about unallocated regions?
> > >>>
> > >>
> > >> You can either send trim commands, or just skip those regions and let
> >
> >>> If we use nbd, how can we pass additional information to the other
> >>> side, for example information about unallocated regions?
> >>>
> >>
> >> You can either send trim commands, or just skip those regions and let
> >> the other side figure it out.
> >
> > But we are lost if we want to tran
On Sat, Feb 23, 2013 at 09:05:37AM +, Dietmar Maurer wrote:
> > > Sure. nbd+unix:///exportname?socket=path is the new URI syntax, I
> > > honestly forgot the old one. SCM_CREDENTIALS checks (qemu-nbd --pid
> > > or something like that) is not supported, but patches would be very
> > > welcom
Il 25/02/2013 13:59, Dietmar Maurer ha scritto:
unix sockets works with qemu nbd code?
>>>
>>> Sure. nbd+unix:///exportname?socket=path is the new URI syntax, I
>>> honestly forgot the old one. SCM_CREDENTIALS checks (qemu-nbd --pid
>>> or something like that) is not supported, but patches w
> >> unix sockets works with qemu nbd code?
> >
> > Sure. nbd+unix:///exportname?socket=path is the new URI syntax, I
> > honestly forgot the old one. SCM_CREDENTIALS checks (qemu-nbd --pid
> > or something like that) is not supported, but patches would be very welcome.
>
> Yes, this is better t
> > So if I want to use the code from nbd.c, I need to write a specialized
> > BlockDriver for the vma format (to pass that to nbd_export_new())?
>
> Yes. But I believe that would be a good thing to do anyway. For one thing,
> it
> gives you "automatic" coverage via qemu-iotests.
But the only
> > > Sure. nbd+unix:///exportname?socket=path is the new URI syntax, I
> > > honestly forgot the old one. SCM_CREDENTIALS checks (qemu-nbd --pid
> > > or something like that) is not supported, but patches would be
> > > very welcome.
> >
> > Yes, this is better than my tcp suggestion.
>
> So i
> -Original Message-
> From: Paolo Bonzini [mailto:pbonz...@redhat.com]
> Sent: Freitag, 22. Februar 2013 19:07
> To: Dietmar Maurer
> Cc: Stefan Hajnoczi; Kevin Wolf; qemu-devel@nongnu.org; Wenchao Xia
> Subject: Re: [PATCH v4 0/6] Efficient VM backup for qemu
>
> Il 22/02/2013 18:57, D
> > Sure. nbd+unix:///exportname?socket=path is the new URI syntax, I
> > honestly forgot the old one. SCM_CREDENTIALS checks (qemu-nbd --pid
> > or something like that) is not supported, but patches would be very welcome.
>
> Yes, this is better than my tcp suggestion.
So if I want to use the
Il 22/02/2013 18:57, Dietmar Maurer ha scritto:
>>> If we use nbd, how can we pass additional information to the other
>>> side, for example information about unallocated regions?
>>>
>>
>> You can either send trim commands, or just skip those regions and let the
>> other
>> side figure it out.
>
> > If we use nbd, how can we pass additional information to the other
> > side, for example information about unallocated regions?
> >
>
> You can either send trim commands, or just skip those regions and let the
> other
> side figure it out.
But we are lost if we want to transfer something els
Il 22/02/2013 18:49, Dietmar Maurer ha scritto:
> >> unix sockets works with qemu nbd code?
>>> > >
>>> > > Sure. nbd+unix:///exportname?socket=path is the new URI syntax, I
>>> > > honestly forgot the old one. SCM_CREDENTIALS checks (qemu-nbd --pid
>>> > > or something like that) is not sup
> On Fri, Feb 22, 2013 at 2:07 PM, Paolo Bonzini wrote:
> > Il 22/02/2013 14:02, Dietmar Maurer ha scritto:
> >>> But you can always sandbox using SELinux, if you care about that, or
> >>> use a Unix socket + SCM_CREDENTIALS.
> >>
> >> unix sockets works with qemu nbd code?
> >
> > Sure. nbd+uni
On Fri, Feb 22, 2013 at 2:07 PM, Paolo Bonzini wrote:
> Il 22/02/2013 14:02, Dietmar Maurer ha scritto:
>>> But you can always sandbox using SELinux, if you care about that, or use a
>>> Unix
>>> socket + SCM_CREDENTIALS.
>>
>> unix sockets works with qemu nbd code?
>
> Sure. nbd+unix:///exportn
On Fri, Feb 22, 2013 at 11:39 AM, Kevin Wolf wrote:
> Am 22.02.2013 um 11:31 hat Stefan Hajnoczi geschrieben:
>> On Thu, Feb 21, 2013 at 03:56:15PM +, Dietmar Maurer wrote:
>> > > On Wed, Feb 20, 2013 at 04:04:49PM +, Dietmar Maurer wrote:
>> I added the 'wait-for' attribute which tells QM
Il 22/02/2013 14:02, Dietmar Maurer ha scritto:
>
> Why is this needed?
>>> Security? I don't want that another process can write nonsense into my
>> backup.
>>
>> They can already write nonsense to your iSCSI target, can't they?
>
> I am more concerned about software bugs. You need to find a
> >> > Why is this needed?
> > Security? I don't want that another process can write nonsense into my
> backup.
>
> They can already write nonsense to your iSCSI target, can't they?
I am more concerned about software bugs. You need to find a free port, and
then pass that port to kvm. If the ori
Il 22/02/2013 12:51, Dietmar Maurer ha scritto:
>> You can also use the named-export support. Something like
>> >
>> >--nbd=127.0.0.1:1235
>> >--blockdev=virtio-drive0 --blockdev=virtio-drive1 --blockdev=ide0-hd0
>> >
>> > etc.
>> >
>> > Then QEMU can migrate to nbd://127.0.0.1:1235/vir
> You can also use the named-export support. Something like
>
>--nbd=127.0.0.1:1235
>--blockdev=virtio-drive0 --blockdev=virtio-drive1 --blockdev=ide0-hd0
>
> etc.
>
> Then QEMU can migrate to nbd://127.0.0.1:1235/virtio-drive0.
>
> > Or how can we guarantee that only one kvm process c
Il 22/02/2013 12:13, Dietmar Maurer ha scritto:
> Opening several NBD sockets 'nbd:127.0.0.1:1235' seems dangerous to me. One
> backup task
> and easily write to the wrong port - there is no protection?
You can also use the named-export support. Something like
--nbd=127.0.0.1:1235
--block
Opening several NBD sockets 'nbd:127.0.0.1:1235' seems dangerous to me. One
backup task
and easily write to the wrong port - there is no protection? Or how can we
guarantee that
only one kvm process can write to that NBD device?
> transaction actions=[{'type': 'migrate-async',
>
Am 22.02.2013 um 11:31 hat Stefan Hajnoczi geschrieben:
> On Thu, Feb 21, 2013 at 03:56:15PM +, Dietmar Maurer wrote:
> > > On Wed, Feb 20, 2013 at 04:04:49PM +, Dietmar Maurer wrote:
> > > > > The backup writer abstraction is a special case interface somewhere
> > > > > between an image fo
On Thu, Feb 21, 2013 at 03:56:15PM +, Dietmar Maurer wrote:
> > On Wed, Feb 20, 2013 at 04:04:49PM +, Dietmar Maurer wrote:
> > > > The backup writer abstraction is a special case interface somewhere
> > > > between an image format and live migration. I'd prefer it if the
> > > > backup bl
Il 21/02/2013 16:56, Dietmar Maurer ha scritto:
>> > In _your_ use case. That means that we should support using non-seekable
>> > pipes, but it doesn't mean that any other use case is irrelevant. In fact
>> > I think
>> > backing up to a normal raw or qcow2 image file is the interesting case for
> On Wed, Feb 20, 2013 at 04:04:49PM +, Dietmar Maurer wrote:
> > > The backup writer abstraction is a special case interface somewhere
> > > between an image format and live migration. I'd prefer it if the
> > > backup block job stuck to BlockDriverState as the destination for backup
> data.
On Wed, Feb 20, 2013 at 04:04:49PM +, Dietmar Maurer wrote:
> > The backup writer abstraction is a special case interface somewhere
> > between an image format and live migration. I'd prefer it if the backup
> > block job stuck to BlockDriverState as the destination for backup data.
> > Then y
On Thu, Feb 21, 2013 at 06:53:24AM +, Dietmar Maurer wrote:
> > How would VMA fit in if the backup block job took a BlockDriverState
> > destination and vmsave worked atomically with backup block jobs? You could
> > have QEMU write to NBD server ports that a VMA writer process has open.
> > VM
On Wed, Feb 20, 2013 at 04:04:49PM +, Dietmar Maurer wrote:
> > Interesting series, the backup block job makes sense to me. Regarding
> > efficiency, I think incremental backup is a must, otherwise regular backups
> > using this approach won't really be a win over backing files.
>
> Increment
On Thu, Feb 21, 2013 at 06:53:24AM +, Dietmar Maurer wrote:
> > Interesting series, the backup block job makes sense to me. Regarding
> > efficiency, I think incremental backup is a must,
>
> One can easily implement incremental backup on top of this patches. That is
> why
> I introduced th
> Interesting series, the backup block job makes sense to me. Regarding
> efficiency, I think incremental backup is a must,
One can easily implement incremental backup on top of this patches. That is why
I introduced the BackupDriver abstraction.
> otherwise regular backups using
> this approac
> Interesting series, the backup block job makes sense to me. Regarding
> efficiency, I think incremental backup is a must, otherwise regular backups
> using this approach won't really be a win over backing files.
Incremental backup is something different, not touched by this code.
You can add th
On Wed, Feb 20, 2013 at 10:31:57AM +0100, Dietmar Maurer wrote:
> This series provides a way to efficiently backup VMs.
>
> * Backup to a single archive file
> * Backup contain all data to restore VM (full backup)
> * Do not depend on storage type or image format
> * Avoid use of temporary storage
This series provides a way to efficiently backup VMs.
* Backup to a single archive file
* Backup contain all data to restore VM (full backup)
* Do not depend on storage type or image format
* Avoid use of temporary storage
* store sparse images efficiently
The file docs/backup.txt contains more d
34 matches
Mail list logo