<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body>
<p dir="ltr"><br>
Den 6 jul 2014 08:17 skrev Brad Bendy <brad.bendy@gmail.com>:<br>
><br>
> Is normal behavior a outage when a snapshot occurs? Ive got the rhev<br>
> builds installed but when I take a snap im getting well over two<br>
> minutes of downtime, on a VM with zero usage.</p>
<p dir="ltr">That behavior is expected, engine pauses the VM during snapshot to ensure consistency. Not as sure about the timeframe though; two mins for a idle VM? When I snap, it's a matter if seconds, but that might be your Gluster spooking. We use a traditional
NFS server. Why not set that up and benchmark the difference?</p>
<p dir="ltr">/K</p>
<p dir="ltr">> The VM goes into "pause"<br>
> state then, I guess the snapshot is a copy of the entire VM? I have<br>
> not got a second host up with those rhev build to see the behavior<br>
> when disk migrating<br>
><br>
> Thanks<br>
><br>
> On Sat, Jul 5, 2014 at 7:51 AM, Brad Bendy <brad.bendy@gmail.com> wrote:<br>
> > There we go, sorry about that! Ill give these a test then. Thanks for the help<br>
> ><br>
> > On Sat, Jul 5, 2014 at 7:39 AM, Karli Sjöberg <Karli.Sjoberg@slu.se> wrote:<br>
> >><br>
> >> Den 5 jul 2014 16:22 skrev Brad Bendy <brad.bendy@gmail.com>:<br>
> >><br>
> >><br>
> >>><br>
> >>> Haha, yeah never have been a Fedora fan, and nothing has changed. Is<br>
> >>> the only big feature im missing out on is snapshots? From what I can<br>
> >>> tell, and in my testing, everything else seems to work. Was deploying<br>
> >>> GlusterFS but without the live migration to another host that is<br>
> >>> somewhat defeated.<br>
> >><br>
> >> VM live migration works, live _disk_ migration does not.<br>
> >><br>
> >>> Only way to get that is with RHEL really then?<br>
> >><br>
> >> No, as I earlier pointed out, there is a place you can get the packages you<br>
> >> need for CentOS:<br>
> >> http://jenkins.ovirt.org/view/All/job/qemu-kvm-rhev_create-rpms_el6/lastStableBuild/<br>
> >><br>
> >> You'll have to download- and force install them over the already installed<br>
> >> versions of those packages on all Hosts and then it'll work.<br>
> >><br>
> >> Though, next time there are updates, yum will update from the standard repos<br>
> >> and it just stops working again until you repeat the procedure.<br>
> >><br>
> >> /K<br>
> >><br>
> >>><br>
> >>> On Sat, Jul 5, 2014 at 7:05 AM, Karli Sjöberg <Karli.Sjoberg@slu.se><br>
> >>> wrote:<br>
> >>> ><br>
> >>> > Den 5 jul 2014 15:57 skrev Brad Bendy <brad.bendy@gmail.com>:<br>
> >>> ><br>
> >>> ><br>
> >>> >><br>
> >>> >> If I use Fedora will "everything" work? I had numerous issues, IIRC I<br>
> >>> >> could not even get the ovirtmgmt switch to install and a few other<br>
> >>> >> things. What version of Fedora do you recommend then?<br>
> >>> ><br>
> >>> > None:) We switched long ago to CentOS and have never looked back, even<br>
> >>> > with<br>
> >>> > these issues. Not worth the headache that is Fedora.<br>
> >>> ><br>
> >>> > /K<br>
> >>> ><br>
> >>> >> Ill do another<br>
> >>> >> install and give that a whirl again.<br>
> >>> >><br>
> >>> >> Thanks!<br>
> >>> >><br>
> >>> >> On Fri, Jul 4, 2014 at 10:33 PM, Karli Sjöberg <Karli.Sjoberg@slu.se><br>
> >>> >> wrote:<br>
> >>> >> ><br>
> >>> >> > Den 5 jul 2014 07:04 skrev Brad Bendy <brad.bendy@gmail.com>:<br>
> >>> >> >><br>
> >>> >> >> Hi,<br>
> >>> >> >><br>
> >>> >> >> Ive seeing conflicting info with what version of qemu rpms are<br>
> >>> >> >> needed<br>
> >>> >> >> to do live migration under CentOS. It appears the stock ones will<br>
> >>> >> >> not<br>
> >>> >> >> work and the RHEV ones are required. All the mailing list post I see<br>
> >>> >> >> are from 3-4 months ago, so not sure.<br>
> >>> >> >><br>
> >>> >> >> Im getting VDSGenericException: VDSErrorException: Failed to<br>
> >>> >> >> SnapshotVDS, error = Snapshot failed, code = 48 (Failed with error<br>
> >>> >> >> SNAPSHOT_FAILED and code 48)<br>
> >>> >> >><br>
> >>> >> >> I also saw this thread:<br>
> >>> >> >> http://comments.gmane.org/gmane.linux.centos.general/138593<br>
> >>> >> >><br>
> >>> >> >> Ive been having issues getting those to install, but before I spent<br>
> >>> >> >> to<br>
> >>> >> >> much more time I wanted to really see if I was on the right track.<br>
> >>> >> >><br>
> >>> >> >> Is there a better OS choice? I first started trying with Fedora 19<br>
> >>> >> >> and<br>
> >>> >> >> 20 and has major issues, went to CentOS 6.5 and this is the first<br>
> >>> >> >> and<br>
> >>> >> >> only issue so far ive ran into.<br>
> >>> >> >><br>
> >>> >> >> Thanks!<br>
> >>> >> >> _______________________________________________<br>
> >>> >> >> Users mailing list<br>
> >>> >> >> Users@ovirt.org<br>
> >>> >> >> http://lists.ovirt.org/mailman/listinfo/users<br>
> >>> >> ><br>
> >>> >> > Well, going with Fedora would at least get you the snapshots working,<br>
> >>> >> > if<br>
> >>> >> > I remember correctly, but that's not something you run in production.<br>
> >>> >> > As you<br>
> >>> >> > said, "major issues".<br>
> >>> >> ><br>
> >>> >> > For CentOS, you need "special" versions of certain packages, since<br>
> >>> >> > RedHat wants you to pay for RHEV, they have chosen to cripple the<br>
> >>> >> > standard<br>
> >>> >> > packages so those features won't work:<br>
> >>> >> > http://lists.ovirt.org/pipermail/devel/2014-June/007735.html<br>
> >>> >> ><br>
> >>> >> > And here you can find the packages you need:<br>
> >>> >> ><br>
> >>> >> ><br>
> >>> >> > http://jenkins.ovirt.org/view/All/job/qemu-kvm-rhev_create-rpms_el6/lastStableBuild/<br>
> >>> >> ><br>
> >>> >> > /K<br>
</p>
</body>
</html>