We've been doing some experimenting with gluster, and have built a stand-alone gluster cluster (not managed by oVirt). We've been able to create a storage domain backed by that gluster cluster and run VMs with their disks on that storage. The problem we have is that when we take a gluster node down for updates, maintenance etc. the entire storage domain goes offline in oVirt. Other gluster clients, that is servers connecting directly to the gluster cluster don't seem to notice if one node goes offline. Is anyone else using gluster storage in oVirt that is not managed within oVirt? -- Beau Sapach *System Administrator | Information Technology Services | University of Alberta Libraries* *Phone: 780.492.4181 | Email: Beau.Sapach@ualberta.ca <Beau.Sapach@ualberta.ca>*
On Mon, Dec 11, 2017 at 8:44 PM, Beau Sapach <bsapach@ualberta.ca> wrote:
We've been doing some experimenting with gluster, and have built a stand-alone gluster cluster (not managed by oVirt). We've been able to create a storage domain backed by that gluster cluster and run VMs with their disks on that storage.
The problem we have is that when we take a gluster node down for updates, maintenance etc. the entire storage domain goes offline in oVirt. Other gluster clients, that is servers connecting directly to the gluster cluster don't seem to notice if one node goes offline.
Is anyone else using gluster storage in oVirt that is not managed within oVirt?
Did you set also the backup-volfile-servers mount option?
-- Beau Sapach *System Administrator | Information Technology Services | University of Alberta Libraries* *Phone: 780.492.4181 <(780)%20492-4181> | Email: Beau.Sapach@ualberta.ca <Beau.Sapach@ualberta.ca>*
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
I did exactly the same mistake with my standalone GlusterFS cluster and now need to take down all Storage Domains in order to fix this mistake. Probably, worth to add a few words about this in Installation guide! On Tue, Dec 12, 2017 at 4:52 PM, Simone Tiraboschi <stirabos@redhat.com> wrote:
On Mon, Dec 11, 2017 at 8:44 PM, Beau Sapach <bsapach@ualberta.ca> wrote:
We've been doing some experimenting with gluster, and have built a stand-alone gluster cluster (not managed by oVirt). We've been able to create a storage domain backed by that gluster cluster and run VMs with their disks on that storage.
The problem we have is that when we take a gluster node down for updates, maintenance etc. the entire storage domain goes offline in oVirt. Other gluster clients, that is servers connecting directly to the gluster cluster don't seem to notice if one node goes offline.
Is anyone else using gluster storage in oVirt that is not managed within oVirt?
Did you set also the backup-volfile-servers mount option?
-- Beau Sapach *System Administrator | Information Technology Services | University of Alberta Libraries* *Phone: 780.492.4181 <(780)%20492-4181> | Email: Beau.Sapach@ualberta.ca <Beau.Sapach@ualberta.ca>*
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
We did use the backup-volfile-servers option but still had trouble. We were simply adding all servers in the cluster as backups, is there a best practice that should be followed? On Tue, Dec 12, 2017 at 8:59 AM, Artem Tambovskiy < artem.tambovskiy@gmail.com> wrote:
I did exactly the same mistake with my standalone GlusterFS cluster and now need to take down all Storage Domains in order to fix this mistake. Probably, worth to add a few words about this in Installation guide!
On Tue, Dec 12, 2017 at 4:52 PM, Simone Tiraboschi <stirabos@redhat.com> wrote:
On Mon, Dec 11, 2017 at 8:44 PM, Beau Sapach <bsapach@ualberta.ca> wrote:
We've been doing some experimenting with gluster, and have built a stand-alone gluster cluster (not managed by oVirt). We've been able to create a storage domain backed by that gluster cluster and run VMs with their disks on that storage.
The problem we have is that when we take a gluster node down for updates, maintenance etc. the entire storage domain goes offline in oVirt. Other gluster clients, that is servers connecting directly to the gluster cluster don't seem to notice if one node goes offline.
Is anyone else using gluster storage in oVirt that is not managed within oVirt?
Did you set also the backup-volfile-servers mount option?
-- Beau Sapach *System Administrator | Information Technology Services | University of Alberta Libraries* *Phone: 780.492.4181 <(780)%20492-4181> | Email: Beau.Sapach@ualberta.ca <Beau.Sapach@ualberta.ca>*
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
-- Beau Sapach *System Administrator | Information Technology Services | University of Alberta Libraries* *Phone: 780.492.4181 | Email: Beau.Sapach@ualberta.ca <Beau.Sapach@ualberta.ca>*
The backup-volfile-servers as an additional mount option should handle the case where one of the servers goes down - storage domain should continue to be available. The servers specified for this option can be the servers participating in your volume. For instance, the set of unique servers from the "gluster volume info" command. If even with this mount option, you're facing an issue - please log a bug with gluster mount logs and vdsm logs. thanks sahina On Wed, Dec 13, 2017 at 12:37 AM, Beau Sapach <bsapach@ualberta.ca> wrote:
We did use the backup-volfile-servers option but still had trouble. We were simply adding all servers in the cluster as backups, is there a best practice that should be followed?
On Tue, Dec 12, 2017 at 8:59 AM, Artem Tambovskiy < artem.tambovskiy@gmail.com> wrote:
I did exactly the same mistake with my standalone GlusterFS cluster and now need to take down all Storage Domains in order to fix this mistake. Probably, worth to add a few words about this in Installation guide!
On Tue, Dec 12, 2017 at 4:52 PM, Simone Tiraboschi <stirabos@redhat.com> wrote:
On Mon, Dec 11, 2017 at 8:44 PM, Beau Sapach <bsapach@ualberta.ca> wrote:
We've been doing some experimenting with gluster, and have built a stand-alone gluster cluster (not managed by oVirt). We've been able to create a storage domain backed by that gluster cluster and run VMs with their disks on that storage.
The problem we have is that when we take a gluster node down for updates, maintenance etc. the entire storage domain goes offline in oVirt. Other gluster clients, that is servers connecting directly to the gluster cluster don't seem to notice if one node goes offline.
Is anyone else using gluster storage in oVirt that is not managed within oVirt?
Did you set also the backup-volfile-servers mount option?
-- Beau Sapach *System Administrator | Information Technology Services | University of Alberta Libraries* *Phone: 780.492.4181 <(780)%20492-4181> | Email: Beau.Sapach@ualberta.ca <Beau.Sapach@ualberta.ca>*
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
-- Beau Sapach *System Administrator | Information Technology Services | University of Alberta Libraries* *Phone: 780.492.4181 | Email: Beau.Sapach@ualberta.ca <Beau.Sapach@ualberta.ca>*
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Hi, I just updated almost all storage domains with backup-volfile-servers mount options, the last one remaining is the hosted_storage domain which serves hosted engine VM. I wonder if this domain also needs to be configured with backup-volfile-servers option? If so how to do this - I can't put this domain on maintenance via web UI. Regards, Artem On Wed, Dec 13, 2017 at 9:03 AM, Sahina Bose <sabose@redhat.com> wrote:
The backup-volfile-servers as an additional mount option should handle the case where one of the servers goes down - storage domain should continue to be available. The servers specified for this option can be the servers participating in your volume. For instance, the set of unique servers from the "gluster volume info" command.
If even with this mount option, you're facing an issue - please log a bug with gluster mount logs and vdsm logs.
thanks sahina
On Wed, Dec 13, 2017 at 12:37 AM, Beau Sapach <bsapach@ualberta.ca> wrote:
We did use the backup-volfile-servers option but still had trouble. We were simply adding all servers in the cluster as backups, is there a best practice that should be followed?
On Tue, Dec 12, 2017 at 8:59 AM, Artem Tambovskiy < artem.tambovskiy@gmail.com> wrote:
I did exactly the same mistake with my standalone GlusterFS cluster and now need to take down all Storage Domains in order to fix this mistake. Probably, worth to add a few words about this in Installation guide!
On Tue, Dec 12, 2017 at 4:52 PM, Simone Tiraboschi <stirabos@redhat.com> wrote:
On Mon, Dec 11, 2017 at 8:44 PM, Beau Sapach <bsapach@ualberta.ca> wrote:
We've been doing some experimenting with gluster, and have built a stand-alone gluster cluster (not managed by oVirt). We've been able to create a storage domain backed by that gluster cluster and run VMs with their disks on that storage.
The problem we have is that when we take a gluster node down for updates, maintenance etc. the entire storage domain goes offline in oVirt. Other gluster clients, that is servers connecting directly to the gluster cluster don't seem to notice if one node goes offline.
Is anyone else using gluster storage in oVirt that is not managed within oVirt?
Did you set also the backup-volfile-servers mount option?
-- Beau Sapach *System Administrator | Information Technology Services | University of Alberta Libraries* *Phone: 780.492.4181 <(780)%20492-4181> | Email: Beau.Sapach@ualberta.ca <Beau.Sapach@ualberta.ca>*
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
-- Beau Sapach *System Administrator | Information Technology Services | University of Alberta Libraries* *Phone: 780.492.4181 | Email: Beau.Sapach@ualberta.ca <Beau.Sapach@ualberta.ca>*
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Hello, since I am on gluster since Ovirt 3.6.x I also need to change this mount option. Is it possible to change mount option for next engine start/host reboot without taking the volume offline. Perhaps by modifiying database or config files ? As for now my gluster Entry-Point glusterfs.mgmt.mydomain.de points to multiple IP adresses (all gluster nodes) like so: # host glusterfs.mgmt.mydomain.de glusterfs.mgmt.mydomain.de has address 172.16.252.126 glusterfs.mgmt.mydomain.de has address 172.16.252.124 glusterfs.mgmt.mydomain.de has address 172.16.252.122 glusterfs.mgmt.mydomain.de has address 172.16.252.127 glusterfs.mgmt.mydomain.de has address 172.16.252.125 glusterfs.mgmt.mydomain.de has address 172.16.252.121 glusterfs.mgmt.mydomain.de has address 172.16.252.128 glusterfs.mgmt.mydomain.de has address 172.16.252.123 Furthermore is it possible to add a new "managed" gluster volume from administration, or import an existing gluster so that it becomes manageable by Ovirt ? Since my Volume wasn't created by setup I'm not able to create gluster volumes that are managed by gluster. (See screenshot). All documentation (since 4.1) only covers creating a managed gluster volume with the help of gdeploywhile setting up Ovirt. Bye Am 13.12.2017 um 09:05 schrieb Artem Tambovskiy:
Hi,
I just updated almost all storage domains with backup-volfile-servers mount options, the last one remaining is the hosted_storage domain which serves hosted engine VM. I wonder if this domain also needs to be configured with backup-volfile-servers option? If so how to do this - I can't put this domain on maintenance via web UI.
Regards, Artem
On Wed, Dec 13, 2017 at 9:03 AM, Sahina Bose <sabose@redhat.com <mailto:sabose@redhat.com>> wrote:
The backup-volfile-servers as an additional mount option should handle the case where one of the servers goes down - storage domain should continue to be available. The servers specified for this option can be the servers participating in your volume. For instance, the set of unique servers from the "gluster volume info" command.
If even with this mount option, you're facing an issue - please log a bug with gluster mount logs and vdsm logs.
thanks sahina
On Wed, Dec 13, 2017 at 12:37 AM, Beau Sapach <bsapach@ualberta.ca <mailto:bsapach@ualberta.ca>> wrote:
We did use the backup-volfile-servers option but still had trouble. We were simply adding all servers in the cluster as backups, is there a best practice that should be followed?
On Tue, Dec 12, 2017 at 8:59 AM, Artem Tambovskiy <artem.tambovskiy@gmail.com <mailto:artem.tambovskiy@gmail.com>> wrote:
I did exactly the same mistake with my standalone GlusterFS cluster and now need to take down all Storage Domains in order to fix this mistake. Probably, worth to add a few words about this in Installation guide!
On Tue, Dec 12, 2017 at 4:52 PM, Simone Tiraboschi <stirabos@redhat.com <mailto:stirabos@redhat.com>> wrote:
On Mon, Dec 11, 2017 at 8:44 PM, Beau Sapach <bsapach@ualberta.ca <mailto:bsapach@ualberta.ca>> wrote:
We've been doing some experimenting with gluster, and have built a stand-alone gluster cluster (not managed by oVirt). We've been able to create a storage domain backed by that gluster cluster and run VMs with their disks on that storage.
The problem we have is that when we take a gluster node down for updates, maintenance etc. the entire storage domain goes offline in oVirt. Other gluster clients, that is servers connecting directly to the gluster cluster don't seem to notice if one node goes offline.
Is anyone else using gluster storage in oVirt that is not managed within oVirt?
Did you set also the backup-volfile-servers mount option?
-- Beau Sapach *System Administrator | Information Technology Services | University of Alberta Libraries* *Phone: 780.492.4181 <tel:%28780%29%20492-4181> | Email: Beau.Sapach@ualberta.ca <mailto:Beau.Sapach@ualberta.ca>*
_______________________________________________ Users mailing list 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>
_______________________________________________ Users mailing list Users@ovirt.org <mailto:Users@ovirt.org> http://lists.ovirt.org/mailman/listinfo/users <http://lists.ovirt.org/mailman/listinfo/users>
-- Beau Sapach *System Administrator | Information Technology Services | University of Alberta Libraries* *Phone: 780.492.4181 | Email: Beau.Sapach@ualberta.ca <mailto:Beau.Sapach@ualberta.ca>*
_______________________________________________ Users mailing list 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 http://lists.ovirt.org/mailman/listinfo/users
-- *Ralf Schenk* fon +49 (0) 24 05 / 40 83 70 fax +49 (0) 24 05 / 40 83 759 mail *rs@databay.de* <mailto:rs@databay.de> *Databay AG* Jens-Otto-Krag-Straße 11 D-52146 Würselen *www.databay.de* <http://www.databay.de> Sitz/Amtsgericht Aachen • HRB:8437 • USt-IdNr.: DE 210844202 Vorstand: Ralf Schenk, Dipl.-Ing. Jens Conze, Aresch Yavari, Dipl.-Kfm. Philipp Hermanns Aufsichtsratsvorsitzender: Wilhelm Dohmen ------------------------------------------------------------------------
participants (5)
-
Artem Tambovskiy -
Beau Sapach -
Ralf Schenk -
Sahina Bose -
Simone Tiraboschi