Re: Delete snapshots task hung
Have you checked this thread : https://lists.ovirt.org/pipermail/users/2016-April/039277.html You can switch to postgre user, then 'source /opt/rhn/postgresql10/enable' & then 'psql engine'. As per the thread you can find illegal snapshots via 'select image_group_id,imagestatus from images where imagestatus =4;' And then update them via 'update images set imagestatus =1 where imagestatus = 4 and <other criteria>; commit' Best Regards, Strahil Nikolov On Oct 13, 2019 15:45, Leo David <leoalex@gmail.com> wrote:
Hi Everyone, Im still not being able to start the vms... Could anyone give me an advice on sorign this out ? Still having th "Bad volume specification" error, although the disk is present on the storage. This issue would force me to reinstall a 10 nodes Openshift cluster from scratch, which would not be so funny.. Thanks,
Leo.
On Fri, Oct 11, 2019 at 7:12 AM Strahil <hunter86_bg@yahoo.com> wrote:
Nah... It's done directly on the DB and I wouldn't recommend such action for Production Cluster. I've done it only once and it was based on some old mailing lists.
Maybe someone from the dev can assist?
On Oct 10, 2019 13:31, Leo David <leoalex@gmail.com> wrote:
Thank you Strahil, Could you tell me what do you mean by changing status ? Is this something to be done in the UI ?
Thanks,
Leo
On Thu, Oct 10, 2019, 09:55 Strahil <hunter86_bg@yahoo.com> wrote:
Maybe you can change the status of the VM in order the engine to know that it has to blockcommit the snapshots.
Best Regards, Strahil Nikolov
On Oct 9, 2019 09:02, Leo David <leoalex@gmail.com> wrote:
Hi Everyone, Please let me know if any thoughts or recommandations that could help me solve this issue.. The real bad luck in this outage is that these 5 vms are part on an Openshift deployment, and now we are not able to start it up... Before trying to sort this at ocp platform level by replacing the failed nodes with new vms, I would rather prefer to do it at the oVirt level and have the vms starting since the disks are still present on gluster. Thank you so much !
Leo
-- Best regards, Leo David
Thank you Strahil, I'll proceed with these steps and come back to you. Cheers, Leo On Tue, Oct 15, 2019, 06:45 Strahil <hunter86_bg@yahoo.com> wrote:
Have you checked this thread : https://lists.ovirt.org/pipermail/users/2016-April/039277.html
You can switch to postgre user, then 'source /opt/rhn/postgresql10/enable' & then 'psql engine'.
As per the thread you can find illegal snapshots via '*select image_group_id,imagestatus from images where imagestatus =4;*'
And then update them via '*update images set imagestatus =1 where imagestatus = 4 and <other **criteria>;** commit'*
Best Regards, Strahil Nikolov
On Oct 13, 2019 15:45, Leo David <leoalex@gmail.com> wrote:
Hi Everyone, Im still not being able to start the vms... Could anyone give me an
Still having th "Bad volume specification" error, although the disk is
advice on sorign this out ? present on the storage.
This issue would force me to reinstall a 10 nodes Openshift cluster from scratch, which would not be so funny.. Thanks,
Leo.
On Fri, Oct 11, 2019 at 7:12 AM Strahil <hunter86_bg@yahoo.com> wrote:
Nah... It's done directly on the DB and I wouldn't recommend such action for
Production Cluster.
I've done it only once and it was based on some old mailing lists.
Maybe someone from the dev can assist?
On Oct 10, 2019 13:31, Leo David <leoalex@gmail.com> wrote:
Thank you Strahil, Could you tell me what do you mean by changing status ? Is this
something to be done in the UI ?
Thanks,
Leo
On Thu, Oct 10, 2019, 09:55 Strahil <hunter86_bg@yahoo.com> wrote:
Maybe you can change the status of the VM in order the engine to know
that it has to blockcommit the snapshots.
Best Regards, Strahil Nikolov
On Oct 9, 2019 09:02, Leo David <leoalex@gmail.com> wrote:
Hi Everyone, Please let me know if any thoughts or recommandations that could
help me solve this issue..
The real bad luck in this outage is that these 5 vms are part on an Openshift deployment, and now we are not able to start it up... Before trying to sort this at ocp platform level by replacing the failed nodes with new vms, I would rather prefer to do it at the oVirt level and have the vms starting since the disks are still present on gluster. Thank you so much !
Leo
-- Best regards, Leo David
Thank you for help Strahil, But although there where 4 images with a status 4 in the database, and did the update query on them, same bloody message, and vms wont start. Eventually, ive decided to delete the vms, and do a from scratch installation. Persistance openshift vms are still ok, so i should be able to reuse the volumes somehow. This is why sometimes a subscription is good, when a lack of knowledge on my side is happening. Production systems should not rely on upstreams unless there is a strong understanding of the product. Again, thank you so much for trying helping me out ! Cheers, Leo On Tue, Oct 15, 2019, 07:00 Leo David <leoalex@gmail.com> wrote:
Thank you Strahil, I'll proceed with these steps and come back to you. Cheers,
Leo
On Tue, Oct 15, 2019, 06:45 Strahil <hunter86_bg@yahoo.com> wrote:
Have you checked this thread : https://lists.ovirt.org/pipermail/users/2016-April/039277.html
You can switch to postgre user, then 'source /opt/rhn/postgresql10/enable' & then 'psql engine'.
As per the thread you can find illegal snapshots via '*select image_group_id,imagestatus from images where imagestatus =4;*'
And then update them via '*update images set imagestatus =1 where imagestatus = 4 and <other **criteria>;** commit'*
Best Regards, Strahil Nikolov
On Oct 13, 2019 15:45, Leo David <leoalex@gmail.com> wrote:
Hi Everyone, Im still not being able to start the vms... Could anyone give me an
Still having th "Bad volume specification" error, although the disk is
advice on sorign this out ? present on the storage.
This issue would force me to reinstall a 10 nodes Openshift cluster from scratch, which would not be so funny.. Thanks,
Leo.
On Fri, Oct 11, 2019 at 7:12 AM Strahil <hunter86_bg@yahoo.com> wrote:
Nah... It's done directly on the DB and I wouldn't recommend such action for
Production Cluster.
I've done it only once and it was based on some old mailing lists.
Maybe someone from the dev can assist?
On Oct 10, 2019 13:31, Leo David <leoalex@gmail.com> wrote:
Thank you Strahil, Could you tell me what do you mean by changing status ? Is this
something to be done in the UI ?
Thanks,
Leo
On Thu, Oct 10, 2019, 09:55 Strahil <hunter86_bg@yahoo.com> wrote:
Maybe you can change the status of the VM in order the engine to
know that it has to blockcommit the snapshots.
Best Regards, Strahil Nikolov
On Oct 9, 2019 09:02, Leo David <leoalex@gmail.com> wrote:
> > Hi Everyone, > Please let me know if any thoughts or recommandations that could help me solve this issue.. > The real bad luck in this outage is that these 5 vms are part on an Openshift deployment, and now we are not able to start it up... > Before trying to sort this at ocp platform level by replacing the failed nodes with new vms, I would rather prefer to do it at the oVirt level and have the vms starting since the disks are still present on gluster. > Thank you so much ! > > > Leo
-- Best regards, Leo David
participants (2)
-
Leo David -
Strahil