Re: [ovirt-users] How long can a disk snapshot exist for?
by Vered Volansky
Hi Ryan,
Should have replied to all, my bad.
See my answer embedded below:
----- Original Message -----
> From: "Ryan Groten" <Ryan.Groten(a)stantec.com>
> To: "Vered Volansky" <vered(a)redhat.com>
> Sent: Thursday, August 28, 2014 5:50:12 PM
> Subject: RE: [ovirt-users] How long can a disk snapshot exist for?
>
> Thanks for the reply! So when keeping a snapshot for a long time I have to
> keep an eye on how large it will get over time.
The snapshot's size itself is determined when it's taken according to the disks size at the time.
It then stays the same.
When taking a snapshot, the active image of the disk is "frozen" at this point in time, and a new, empty active image is created to hold the new data on the disk.
The new data is saved in the form of diffs, so if there are mainly additions to the snapshot time, the space difference of the snapshot is negligible.
But if the diffs include many reductions from the snapshot's state - this might consume a lot of space, again, depending on your usage.
Note that you're also limited by the disk's size.
But there's no (or very
> little) performance impact or potential issues from keeping snapshots (other
> than the storage pool filling up maybe)?
All the vm operations take into consideration snapshots. Storage is an issue, but so is every operation you'll make on the vm.
When you have one image, the vm is handled only through this image. But when you have several, for each operation the right layer will search for the existence/ability to do the operation in all the snapshots (worst case).
So there is in fact an impact, but it's due to the mere existence of the snapshots and their number, not their age.
We support 26 snapshots per VM, but you should only use it if you actually need the backup.
If you need RT performance, try to avoid it as possible.
>
> Thanks,
> Ryan
>
> -----Original Message-----
> From: Vered Volansky [mailto:vered@redhat.com]
> Sent: August-27-14 11:19 PM
> To: Groten, Ryan
> Subject: Re: [ovirt-users] How long can a disk snapshot exist for?
>
> Ryan,
>
> Disk snapshots consume fixed storage space (fixed since time of creation).
> The more differences there are from on your disk since the snapshot was
> taken, the more space is consumed, but that happens with no relation to the
> snapshot.
> If your frequent changes are in the form of adding data to you disks (on top
> of data snapshot time), then the space consumption of the snapshot is
> negligible.
> If you are undoing stuff from the snapshot time, there is actually more space
> consumed (to save the differences), otherwise the space would have just been
> released.
>
> Vered
>
>
> ----- Original Message -----
> > From: "Ryan Groten" <Ryan.Groten(a)stantec.com>
> > To: users(a)ovirt.org
> > Sent: Thursday, August 28, 2014 12:49:02 AM
> > Subject: [ovirt-users] How long can a disk snapshot exist for?
> >
> >
> >
> > Is there any limit/performance considerations to keeping a disk
> > snapshot for extended periods of time? What if the disk is changing
> > frequently vs mostly static?
> >
> >
> >
> > Thanks,
> >
> > Ryan
> >
> > _______________________________________________
> > Users mailing list
> > Users(a)ovirt.org
> > http://lists.ovirt.org/mailman/listinfo/users
> >
>
10 years, 10 months
oVirt/gluster storage questions for 2-3 node datacenter
by David King
Hi,
I am currently testing oVirt 3.4.3 + gluster 3.5.2 for use in my relatively
small home office environment on a single host. I have 2 Intel hosts with
SSD and magnetic disk and one AMD host with only magnetic disk. I have
been trying to figure out the best way to configure my environment given my
previous attempt with oVirt 3.3 encountered storage issues.
I will be hosting two types of VMs - VMs that can be tied to a particular
system (such as 3 node FreeIPA domain or some test VMs), and VMs which
could migrate between systems for improved uptime.
The processor issue seems straightforward. Have a single datacenter with
two clusters - one for the Intel systems and one for the AMD systems. Put
VMs which need to live migrate on the Intel cluster. If necessary VMs can
be manually switched between the Intel and AMD cluster with a downtime.
The Gluster side of the storage seems less clear. The bulk of the gluster
with oVirt issues I experienced and have seen on the list seem to be two
node setups with 2 bricks in the Gluster volume.
So here are my questions:
1) Should I avoid 2 brick Gluster volumes?
2) What is the risk in having the SSD volumes with only 2 bricks given that
there would be 3 gluster servers? How should I configure them?
3) Is there a way to use local storage for a host locked VM other than
creating a gluster volume with one brick?
4) Should I avoid using the hosted engine configuration? I do have an
external VMware ESXi system to host the engine for now but would like to
phase it out eventually.
5) If I do the hosted engine should I make the underlying gluster volume 3
brick replicated?
Thanks in advance for any help you can provide.
-David
10 years, 10 months
Import Windows VHD disk to oVirt virtio-scsi dsik in Admin portl
by Grzegorz Szypa
Hi.
Anyone have any idea how to do it.
Normally I did it this way that:
- Convert VHD format IMG:
Example: "qemu-img convert -f vpc -O raw VHD_WinSrvRAS2012.vhd
/usr/local/VHD_WinSrvRAS2012.img -p"
- I was copying the file RAW disk:
Example: "dd if=VHD_WinSrvRAS2012.img
of=/usr/local/virtuos/99a241af-2a65-468b-8df2-
fa52c430ed5e/images/15f65c23-319e-4b9b-8313-14261886f7cd/4ffd8834-b99d-46d7-990b-a73c3e66a30e
"
But when I run the machine there is a problem with the detection of the
disk operating system. I was wondering is there a way to tell him to add
drivers for virtio-scsi interface. But i do not know how to do it?
--
G.Sz.
10 years, 10 months
Functionalities of GlusterFS (vs VMFS)
by David BERCOT
Hello,
Maybe I'm border-line with oVirt but I think it is a critical part of a
virtual infrastructure, so...
I understand that GlusterFS is a good storage solution for oVirt VM but
I have a couple of questions :
- does GlusterFS support FibreChannel (in order to have equivalents
to the SAN Datastores of the VMware world) ?
- is GlusterFS performant enough to support a lot of VM with a lot of
nodes (entreprise infrastructure) ?
- has GlusterFS some limitations or constraints ?
- can GlusterFS be used (out of oVirt) for high I/O needs ?
Thank you.
David.
10 years, 10 months
Snapshot preview,clone and undo using ovirt api
by Shanil S
Hi All,
In the ovirt admin panel, it is possible to do the following snapshots
actions
1. Clone
2. Preview
3. Undo
but i couldn't find out any direct api methods to do these actions, Is it
possible to it using the ovirt api ?
--
Regards
Shanil
10 years, 10 months
Re: [ovirt-users] Problem starting VM in my ovirt enviroment
by Andy Michielsen
Hello Dan,
Nevermind. After 2 months I had some time to look at this problem again but
didn't remember I put the server in maintenance mode.
I put it in maintenance mode to install updates and doublecheck all
configuration and apparently I did something wright as it just works now.
Kind regards.
2014-08-28 16:34 GMT+02:00 Andy Michielsen <andy.michielsen(a)gmail.com>:
> Hello Dan,
>
> I can't find a clue in the engine.log but I'm unable to start any VM on
> this host.
> Looks like it lost its connection to its datadomain entirely.
>
> Kind regards.
>
>
> 2014-08-28 15:12 GMT+02:00 Dan Kenigsberg <danken(a)redhat.com>:
>
> On Thu, Aug 28, 2014 at 02:14:07PM +0200, Andy Michielsen wrote:
>> > Dear all,
>> >
>> > I need some help. I wanted to start a VM but got the following error.
>> >
>> > [image: Inline afbeelding 1]
>> >
>> > What's going on here ? How can I fix this.
>>
>> does engine.log have any clues regarding the falure to find a valid
>> host? Do you have guess (are they all operational? are they over-worked?
>>
>
>
10 years, 10 months
Problem starting VM in my ovirt enviroment
by Andy Michielsen
Dear all,
I need some help. I wanted to start a VM but got the following error.
[image: Inline afbeelding 1]
What's going on here ? How can I fix this.
Kind regards.
10 years, 10 months
Prestarted vms in a pool
by Michel Rode
--lPJ5i9rX1WvdYcWF
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable
Hi all,
we have a little problem here...we created a new pool with 10 vms and 4
of them as prestarted.
During the sysprep (Windows vms) the vms getting rebooted and this is
the problem:
2014-08-27 17:22:47,759 INFO
[org.ovirt.engine.core.vdsbroker.VdsUpdateRunTimeInfo]
(DefaultQuartzScheduler_Worker-54) VM PC53VP0002-3 18d06e9b-3674-4ad6-b3
c9-511416f69ae0 moved from Up --> RebootInProgress
2014-08-27 17:22:47,775 INFO
[org.ovirt.engine.core.vdsbroker.VdsUpdateRunTimeInfo]
(DefaultQuartzScheduler_Worker-54) VM PC53VP0002-2 77a3e6ca-bcaf-4415-a0
20-34e0f43307d4 moved from Up --> RebootInProgress
2014-08-27 17:22:47,785 INFO
[org.ovirt.engine.core.vdsbroker.VdsUpdateRunTimeInfo]
(DefaultQuartzScheduler_Worker-54) VM PC53VP0002-4 3430c339-cc89-45e5-af
e6-49b217b2598d moved from Up --> RebootInProgress
2014-08-27 17:22:47,846 INFO
[org.ovirt.engine.core.vdsbroker.VdsUpdateRunTimeInfo]
(DefaultQuartzScheduler_Worker-75) VM PC53VP0002-1 825850ef-e63c-4c99-93
89-400f1af22e5f moved from Up --> RebootInProgress
2014-08-27 17:22:57,027 INFO [org.ovirt.engine.core.bll.VmPoolMonitor]
(DefaultQuartzScheduler_Worker-60) [6e7094ae] VmPool
4f5fe15b-123c-4beb-b596-528747f2
a6ec is missing 4 prestarted Vms, attempting to prestart 4 Vms
And after 10mins all vms in the pool are started!
Why is VmPoolMonitor attempting to prestart 4 Vms again?
Thanks!
Michel
--=20
Michel Rode =20
Linux / Unix Consultant & Trainer=20
Mail: rode(a)b1-systems.de
B1 Systems GmbH
Osterfeldstra=DFe 7 / 85088 Vohburg / http://www.b1-systems.de
GF: Ralph Dehner / Unternehmenssitz: Vohburg / AG: Ingolstadt,HRB 3537
--lPJ5i9rX1WvdYcWF
Content-Type: application/pgp-signature
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
iEYEARECAAYFAlP+NxwACgkQfKnlh+KjmY1fEQCeOq/Rxx5JeiPTIyuS62tBAioH
PoIAnAyu42ip1Kxbl0xoxi3Jx0JNFwUO
=cUZx
-----END PGP SIGNATURE-----
--lPJ5i9rX1WvdYcWF--
10 years, 10 months
Moving a POSIX storage domain to be a Local Storage
by Elad Shmitanka
Hi,
In the initial setup I wasn't aware of the option to assign a local disk as
Local Storage and we have added it as a POSIX complaint Storage Domain, now
that we want to add a new KVM host, it needs to attach the storage domain
to the new host, which it obviously can't.
My question is, how do I remove the Storage domain and add it back as a
Local storage without harming the VMs that are already on the system
(obviously some downtime will be needed)
I guess that I can add another Storage Domain and start moving them to a
"temp location" and then back, question is if there is another quicker way.
Thanks
--
Elad Shmitanka
DevOps Engineer
Mobile: *+972-52-2259939* | *elad(a)myheritage.com <elad(a)myheritage.com>*
Tel: *+972-3-6280000 Ext. 105* | www.myheritage.com
U.S.A Number: 17183608065 Ext. 2
MyHeritage Ltd., 3 Ariel Sharon St., Terminal Park, Or Yehuda 60250, Israel
<http://www.myheritage.com/> <http://blog.myheritage.com/>
<https://www.facebook.com/myheritage> <https://twitter.com/myheritage>
10 years, 10 months
iSCSI self hosted engine
by Markus Mathes
Hi,
I thought about a self hosted engine setup using 2 hosts connected to
a iSCSI storage.
A few questions arose and I didn't succeed finding the answers on the
web. I somehow don't get how the iSCSI integration in ovirt works.
- As far as I understand from the information I found, all iSCSI
access for the virtual machines is done through one host. Does this
mean, that all disk related traffic of the other host is going to the
first host and then to the iSCSI storage? In this case. How are the
virtual disk images then made available for the VMs on the other host?
- If the first host fails, the engine will get started on the other
host. Will the engine deal with the fact, that now the iSCSI storage
domain has to be accessed using the other host. Or is iSCSI just not
suitable for this kind of usage.
- Has the self hosted engine to use a different iSCSI lun than the
storage domain?
Thanks
Markus
--
**************************************************
Markus Mathes, PhD
Product Development
www.dectris.com
**************************************************
10 years, 10 months
