glusterfs resume vm paused state

Hi, Has anyone had any luck with resuming a VM from a paused state on top of NFS share? Even when the VMs are marked as HA, if the gluster storage goes down for a few seconds the VMs go to a paused state and can never be resumed. They require a hard reset. I recall when using NFS to not have this issue. Thanks, Andrew

Hi Andrew, Afaict, there should be manual intervention to resume a 'paused vm' in any storage domain even if VM is marked as "HA".. Also, I failed to understand the setup you have, that said, you mentioned: " resuming a VM from a paused state on top of NFS share? Even when the VMs are marked as HA, if the gluster storage goes down for a few seconds the VMs go to a paused state and can never be resumed" Do you have NFS storage domain configured by specifying "gluster server ip" and "volume name " in place of "server" and "export" path ? can you please detail the setup (wrt storage domain configuration and gluster volumes) and version of ovirt and gluster in use ? --Humble On Mon, Jun 2, 2014 at 11:17 AM, Andrew Lau <andrew@andrewklau.com> wrote:
Hi,
Has anyone had any luck with resuming a VM from a paused state on top of NFS share? Even when the VMs are marked as HA, if the gluster storage goes down for a few seconds the VMs go to a paused state and can never be resumed. They require a hard reset.
I recall when using NFS to not have this issue.
Thanks, Andrew _______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users

Hi Humble, On Mon, Jun 2, 2014 at 8:10 PM, Humble Devassy Chirammal <humble.devassy@gmail.com> wrote:
Hi Andrew,
Afaict, there should be manual intervention to resume a 'paused vm' in any storage domain even if VM is marked as "HA"..
I had a BZ open about this with some traction, but i forgot to keep up with the requests and it's fallen behind https://bugzilla.redhat.com/show_bug.cgi?id=1058300 Even manually, they won't resume. virsh resume host also has the same end result.
Also, I failed to understand the setup you have, that said, you mentioned:
" resuming a VM from a paused state on top of NFS share? Even when the VMs are marked as HA, if the gluster storage goes down for a few seconds the VMs go to a paused state and can never be resumed"
Do you have NFS storage domain configured by specifying "gluster server ip" and "volume name " in place of "server" and "export" path ?
can you please detail the setup (wrt storage domain configuration and gluster volumes) and version of ovirt and gluster in use ?
We're testing a two host setup with oVirt and gluster on the same boxes. CentOS 6.5, hosted-engine. Storage domain type as glusterfs, although when I try a storage domain type of nfs (using the gluster nfs server) the above issue doesn't seem to occur.
--Humble
On Mon, Jun 2, 2014 at 11:17 AM, Andrew Lau <andrew@andrewklau.com> wrote:
Hi,
Has anyone had any luck with resuming a VM from a paused state on top of NFS share? Even when the VMs are marked as HA, if the gluster storage goes down for a few seconds the VMs go to a paused state and can never be resumed. They require a hard reset.
I recall when using NFS to not have this issue.
Thanks, Andrew _______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users

On 06/02/2014 01:10 PM, Humble Devassy Chirammal wrote:
Hi Andrew,
Afaict, there should be manual intervention to resume a 'paused vm' in any storage domain even if VM is marked as "HA"..
that has been fixed since 3.3 with auto-resume paused vm's after EIO
Also, I failed to understand the setup you have, that said, you mentioned:
" resuming a VM from a paused state on top of NFS share? Even when the VMs are marked as HA, if the gluster storage goes down for a few seconds the VMs go to a paused state and can never be resumed"
Do you have NFS storage domain configured by specifying "gluster server ip" and "volume name " in place of "server" and "export" path ?
can you please detail the setup (wrt storage domain configuration and gluster volumes) and version of ovirt and gluster in use ?
--Humble
On Mon, Jun 2, 2014 at 11:17 AM, Andrew Lau <andrew@andrewklau.com <mailto:andrew@andrewklau.com>> wrote:
Hi,
Has anyone had any luck with resuming a VM from a paused state on top of NFS share? Even when the VMs are marked as HA, if the gluster storage goes down for a few seconds the VMs go to a paused state and can never be resumed. They require a hard reset.
I recall when using NFS to not have this issue.
Thanks, Andrew _______________________________________________ Users mailing list Users@ovirt.org <mailto:Users@ovirt.org> http://lists.ovirt.org/mailman/listinfo/users
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users

Hi,
that has been fixed since 3.3 with auto-resume paused vm's after EIO
Thanks Itamar .
I had a BZ open about this with some traction, but i forgot to keep up with the requests and it's fallen behind https://bugzilla.redhat.com/show_bug.cgi?id=1058300 Even manually, they won't resume. virsh resume host also has the same end result.
@Andrew , the log files have to be analysed further. Its better to follow up in the bugzilla. --Humble On Tue, Jun 3, 2014 at 5:02 PM, Itamar Heim <iheim@redhat.com> wrote:
On 06/02/2014 01:10 PM, Humble Devassy Chirammal wrote:
Hi Andrew,
Afaict, there should be manual intervention to resume a 'paused vm' in any storage domain even if VM is marked as "HA"..
that has been fixed since 3.3 with auto-resume paused vm's after EIO
Also, I failed to understand the setup you have, that said, you mentioned:
" resuming a VM from a paused state on top of NFS share? Even when the VMs are marked as HA, if the gluster storage goes down for a few seconds the VMs go to a paused state and can never be resumed"
Do you have NFS storage domain configured by specifying "gluster server ip" and "volume name " in place of "server" and "export" path ?
can you please detail the setup (wrt storage domain configuration and gluster volumes) and version of ovirt and gluster in use ?
--Humble
On Mon, Jun 2, 2014 at 11:17 AM, Andrew Lau <andrew@andrewklau.com <mailto:andrew@andrewklau.com>> wrote:
Hi,
Has anyone had any luck with resuming a VM from a paused state on top of NFS share? Even when the VMs are marked as HA, if the gluster storage goes down for a few seconds the VMs go to a paused state and can never be resumed. They require a hard reset.
I recall when using NFS to not have this issue.
Thanks, Andrew _______________________________________________ Users mailing list Users@ovirt.org <mailto:Users@ovirt.org> http://lists.ovirt.org/mailman/listinfo/users
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users

On 06/16/2014 04:20 PM, Humble Devassy Chirammal wrote:
Hi,
that has been fixed since 3.3 with auto-resume paused vm's after EIO
Thanks Itamar .
I had a BZ open about this with some traction, but i forgot to keep up with the requests and it's fallen behind https://bugzilla.redhat.com/show_bug.cgi?id=1058300
i assume this is around gluster deployment/split-brains/etc., since we try to resume and fail
Even manually, they won't resume. virsh resume host also has the same end result.
@Andrew , the log files have to be analysed further. Its better to follow up in the bugzilla.
--Humble
On Tue, Jun 3, 2014 at 5:02 PM, Itamar Heim <iheim@redhat.com <mailto:iheim@redhat.com>> wrote:
On 06/02/2014 01:10 PM, Humble Devassy Chirammal wrote:
Hi Andrew,
Afaict, there should be manual intervention to resume a 'paused vm' in any storage domain even if VM is marked as "HA"..
that has been fixed since 3.3 with auto-resume paused vm's after EIO
Also, I failed to understand the setup you have, that said, you mentioned:
" resuming a VM from a paused state on top of NFS share? Even when the VMs are marked as HA, if the gluster storage goes down for a few seconds the VMs go to a paused state and can never be resumed"
Do you have NFS storage domain configured by specifying "gluster server ip" and "volume name " in place of "server" and "export" path ?
can you please detail the setup (wrt storage domain configuration and gluster volumes) and version of ovirt and gluster in use ?
--Humble
On Mon, Jun 2, 2014 at 11:17 AM, Andrew Lau <andrew@andrewklau.com <mailto:andrew@andrewklau.com> <mailto:andrew@andrewklau.com <mailto:andrew@andrewklau.com>>__> wrote:
Hi,
Has anyone had any luck with resuming a VM from a paused state on top of NFS share? Even when the VMs are marked as HA, if the gluster storage goes down for a few seconds the VMs go to a paused state and can never be resumed. They require a hard reset.
I recall when using NFS to not have this issue.
Thanks, Andrew _________________________________________________ Users mailing list Users@ovirt.org <mailto:Users@ovirt.org> <mailto:Users@ovirt.org <mailto:Users@ovirt.org>> http://lists.ovirt.org/__mailman/listinfo/users <http://lists.ovirt.org/mailman/listinfo/users>
_________________________________________________ Users mailing list Users@ovirt.org <mailto:Users@ovirt.org> http://lists.ovirt.org/__mailman/listinfo/users <http://lists.ovirt.org/mailman/listinfo/users>

Hi Itamar,
i assume this is around gluster deployment/split-brains/etc., since we try to resume and fail
We are progressing on bz. Looks like quorum configuration playing a role here. I will follow up on the same bug report. --Humble On Mon, Jun 16, 2014 at 7:10 PM, Itamar Heim <iheim@redhat.com> wrote:
On 06/16/2014 04:20 PM, Humble Devassy Chirammal wrote:
Hi,
that has been fixed since 3.3 with auto-resume paused vm's after EIO
Thanks Itamar .
I had a BZ open about this with some traction, but i forgot to keep up with the requests and it's fallen behind https://bugzilla.redhat.com/show_bug.cgi?id=1058300
i assume this is around gluster deployment/split-brains/etc., since we try to resume and fail
Even manually, they won't resume. virsh resume host also has the same end result.
@Andrew , the log files have to be analysed further. Its better to follow up in the bugzilla.
--Humble
On Tue, Jun 3, 2014 at 5:02 PM, Itamar Heim <iheim@redhat.com <mailto:iheim@redhat.com>> wrote:
On 06/02/2014 01:10 PM, Humble Devassy Chirammal wrote:
Hi Andrew,
Afaict, there should be manual intervention to resume a 'paused vm' in any storage domain even if VM is marked as "HA"..
that has been fixed since 3.3 with auto-resume paused vm's after EIO
Also, I failed to understand the setup you have, that said, you mentioned:
" resuming a VM from a paused state on top of NFS share? Even when the VMs are marked as HA, if the gluster storage goes down for a few seconds the VMs go to a paused state and can never be resumed"
Do you have NFS storage domain configured by specifying "gluster server ip" and "volume name " in place of "server" and "export" path ?
can you please detail the setup (wrt storage domain configuration and gluster volumes) and version of ovirt and gluster in use ?
--Humble
On Mon, Jun 2, 2014 at 11:17 AM, Andrew Lau <andrew@andrewklau.com <mailto:andrew@andrewklau.com> <mailto:andrew@andrewklau.com <mailto:andrew@andrewklau.com>>__>
wrote:
Hi,
Has anyone had any luck with resuming a VM from a paused state on top of NFS share? Even when the VMs are marked as HA, if the gluster storage goes down for a few seconds the VMs go to a paused state and can never be resumed. They require a hard reset.
I recall when using NFS to not have this issue.
Thanks, Andrew _________________________________________________ Users mailing list Users@ovirt.org <mailto:Users@ovirt.org> <mailto:Users@ovirt.org <mailto:Users@ovirt.org>> http://lists.ovirt.org/__mailman/listinfo/users <http://lists.ovirt.org/mailman/listinfo/users>
_________________________________________________
Users mailing list Users@ovirt.org <mailto:Users@ovirt.org> http://lists.ovirt.org/__mailman/listinfo/users <http://lists.ovirt.org/mailman/listinfo/users>
participants (3)
-
Andrew Lau
-
Humble Devassy Chirammal
-
Itamar Heim