You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@cloudstack.apache.org by Valery Ciareszka <va...@gmail.com> on 2013/03/07 18:31:19 UTC
storage issues
Hi all,
I use CS 4.0.1 (1 management node, 1 KVM node, 2 storage nodes that act
like single nfs server (HA) - have primary and secondary storage there, one
more NFS server for another secondary storage (having 2 SS in one zone)
I've faced next issues:
-KVM node tries to reboot when primary storage is offline. I had made wrap
script on /sbin/reboot and found that reboot is called from kvmheartbeat.sh
script:
reboot called at Thu Mar 7 11:08:19 EST 2013
root 9943 0.0 0.0 108164 1528 ? S 11:08 0:00 /bin/bash
/usr/lib64/cloud/common/scripts/vm/hypervisor/kvm/kvmheartbeat.sh -i
10.6.10.1 -p /GIGO/cloud -m /mnt/50127bd7-d195-3f65-83ea-f131c2d5dc73 -c
I tested how VM feels itself upon offlining primary storage - when primary
storage returns back, everything is ok inside VM.
So I have the question - why KVM node tries to reboot itself ?
Several more questions about secondary storage.
-If I have more than 1 secondary storage configured within one zone, how
does cloudstack choose on which of them it will store
templates/iso/snapshots ?
-I realized that contents of 2 sec storage are different. Imagine we have
lost contents of one of the secondary storage permamently. But we have 2
important templates that were backuped. But they have 100gb+.Is there way
to put them on new secondary storage without uploading 100GB through
webmanagement ?
-Is there way to keep both secondary storages in sync ?
--
Regards,
Valery
http://protocol.by/slayer
Re: storage issues
Posted by Ahmad Emneina <ae...@gmail.com>.
So the rebooting is part of a process called fencing. Fencing, a safety
mechanism, prevents the hosts from trying to write to a resource thats in a
questionable state. The secondary storage questions will have to be fielded
by someone with some more knowledge.
On Thu, Mar 7, 2013 at 9:31 AM, Valery Ciareszka
<va...@gmail.com>wrote:
> Hi all,
>
> I use CS 4.0.1 (1 management node, 1 KVM node, 2 storage nodes that act
> like single nfs server (HA) - have primary and secondary storage there, one
> more NFS server for another secondary storage (having 2 SS in one zone)
> I've faced next issues:
>
> -KVM node tries to reboot when primary storage is offline. I had made wrap
> script on /sbin/reboot and found that reboot is called from kvmheartbeat.sh
> script:
>
> reboot called at Thu Mar 7 11:08:19 EST 2013
> root 9943 0.0 0.0 108164 1528 ? S 11:08 0:00 /bin/bash
> /usr/lib64/cloud/common/scripts/vm/hypervisor/kvm/kvmheartbeat.sh -i
> 10.6.10.1 -p /GIGO/cloud -m /mnt/50127bd7-d195-3f65-83ea-f131c2d5dc73 -c
>
> I tested how VM feels itself upon offlining primary storage - when primary
> storage returns back, everything is ok inside VM.
>
> So I have the question - why KVM node tries to reboot itself ?
>
> Several more questions about secondary storage.
>
> -If I have more than 1 secondary storage configured within one zone, how
> does cloudstack choose on which of them it will store
> templates/iso/snapshots ?
>
> -I realized that contents of 2 sec storage are different. Imagine we have
> lost contents of one of the secondary storage permamently. But we have 2
> important templates that were backuped. But they have 100gb+.Is there way
> to put them on new secondary storage without uploading 100GB through
> webmanagement ?
>
> -Is there way to keep both secondary storages in sync ?
>
>
> --
> Regards,
> Valery
>
> http://protocol.by/slayer
>
Re: storage issues
Posted by Nitin Mehta <Ni...@citrix.com>.
On 07/03/13 11:01 PM, "Valery Ciareszka" <va...@gmail.com> wrote:
>Hi all,
>
>I use CS 4.0.1 (1 management node, 1 KVM node, 2 storage nodes that act
>like single nfs server (HA) - have primary and secondary storage there,
>one
>more NFS server for another secondary storage (having 2 SS in one zone)
>I've faced next issues:
>
>-KVM node tries to reboot when primary storage is offline. I had made wrap
>script on /sbin/reboot and found that reboot is called from
>kvmheartbeat.sh
>script:
>
>reboot called at Thu Mar 7 11:08:19 EST 2013
>root 9943 0.0 0.0 108164 1528 ? S 11:08 0:00 /bin/bash
>/usr/lib64/cloud/common/scripts/vm/hypervisor/kvm/kvmheartbeat.sh -i
>10.6.10.1 -p /GIGO/cloud -m /mnt/50127bd7-d195-3f65-83ea-f131c2d5dc73 -c
>
>I tested how VM feels itself upon offlining primary storage - when primary
>storage returns back, everything is ok inside VM.
>
>So I have the question - why KVM node tries to reboot itself ?
>
>Several more questions about secondary storage.
>
>-If I have more than 1 secondary storage configured within one zone, how
>does cloudstack choose on which of them it will store
>templates/iso/snapshots ?
It is chosen randomly for private templates (you can open an enhancement
to have a space based allocator). Public templates should be copied in all
sec. storage.
Multiple sec. storage is for more disk space and higher availability for
public templates.
>
>-I realized that contents of 2 sec storage are different. Imagine we have
>lost contents of one of the secondary storage permamently. But we have 2
>important templates that were backuped. But they have 100gb+.Is there way
>to put them on new secondary storage without uploading 100GB through
>webmanagement ?
If templates were public they will be available in the other sec. storage
as well. In case of private template you would need to again upload it.
>
>-Is there way to keep both secondary storages in sync ?
As stated above only public templates should be copied in all sec.
storages.
In case you want this behaviour I suggest you opening an enhancement bug
at Jira.
>
>
>--
>Regards,
>Valery
>
>http://protocol.by/slayer