On Tue, Aug 9, 2016 at 8:39 PM, Budur Nagaraju <nbudoor(a)gmail.com&gt; wrote: > HI > > How to upgrade oVirt3.5 to 4 ? am using centos6.7. First upgrade or reinstall your hosts with el7. oVirt 3.6+ does not provide packages for el6. el6 hosts will still work, but easiest to upgrade to el7 in 3.5. If you are using hosted-engine, you can follow [1]. Then upgrade to 3.6 using regular docs [2]. In 4.0, the engine is also el7 only. To upgrade, use engine-backup to backup, then reinstall the machine with el7, then restore. I do not think we have a special page for this, but you can follow the doc text of [3], which is also copied in the 4.0 release notes [4]. Please review [5] for more information about engine backup and restore. If you using hosted-engine, you can use the new upgrade tool to upgrade the engine vm using the engine appliance. Not sure that's properly documented either, see also [6]. Also note that this bug is for 4.0.2, to be released soon.

As always, if at all possible, better try first on a test system, especially if it's in production, and backup everything beforehand, including your VMs etc. Best regards, [1] https://www.ovirt.org/documentation/how-to/hosted-engine-host-OS-upgrade/ [2] http://www.ovirt.org/develop/release-management/releases/3.6/ [3] https://bugzilla.redhat.com/show_bug.cgi?id=1318580 [4] https://www.ovirt.org/release/4.0.0/ [5] http://www.ovirt.org/develop/release-management/features/engine/engine-ba... [6] https://bugzilla.redhat.com/show_bug.cgi?id=1319457 -- Didi _______________________________________________ Users mailing list Users(a)ovirt.org http://lists.ovirt.org/mailman/listinfo/users

On Wed, Aug 10, 2016 at 9:42 AM, Simone Tiraboschi <stirabos(a)redhat.com&gt; wrote: > > On Wed, Aug 10, 2016 at 7:53 AM, Yedidyah Bar David <didi(a)redhat.com&gt; > wrote: > > On Tue, Aug 9, 2016 at 8:39 PM, Budur Nagaraju <nbudoor(a)gmail.com&gt; > > wrote: > >> HI > >> > >> How to upgrade oVirt3.5 to 4 ? am using centos6.7. > > > > First upgrade or reinstall your hosts with el7. > > oVirt 3.6+ does not provide packages for el6. el6 hosts will > > still work, but easiest to upgrade to el7 in 3.5. > > > > If you are using hosted-engine, you can follow [1]. > > > > Then upgrade to 3.6 using regular docs [2]. > > > > In 4.0, the engine is also el7 only. To upgrade, use engine-backup > > to backup, then reinstall the machine with el7, then restore. I do > > not think we have a special page for this, but you can follow the > > doc text of [3], which is also copied in the 4.0 release notes [4]. > > Please review [5] for more information about engine backup and restore. > > If you using hosted-engine, you can use the new upgrade tool to upgrade > > the engine vm using the engine appliance. Not sure that's properly > > documented either, see also [6]. Also note that this bug is for 4.0.2, > > to be released soon. > > We have this: > > https://www.ovirt.org/develop/release-management/features/hosted-engine-m... > > But I have not yet understood, based on this thread in June and the related bugzillas: http://lists.ovirt.org/pipermail/users/2016-June/040587.html what is the current situation in 4.0.1 and then 4.0.2 for the problem of upgrading cluster level of the cluster where engine vm is running, in case of Self Hosted Engine environments. And in case not feasible yet, if there is any supported workaround doing modifications at engine database level. Can you give an update on this?

On Wed, Aug 10, 2016 at 11:52 AM, Yedidyah Bar David <didi(a)redhat.com&gt; wrote: > > > > > > Can you give an update on this? > > Should be fixed in 4.0.2: > > https://bugzilla.redhat.com/show_bug.cgi?id=1348907 > > As a general note, if you see a bug on 4.0.z in MODIFIED or QA, > the fix is already in the nightly 4.0 repos, in case you feel like trying: > > http://www.ovirt.org/develop/dev-process/install-nightly-snapshot/ > > Best, > -- > Didi Ok, thanks for the update and clarifications. From an operational point of view, what will change? RIght now, in 4.0, if I select Default cluster and try to edit it and upgrade its Compatibility Version to 4.0, it seems I don't get any message in the gui, or at least I don't see it; it's something similar to when you have sort of syntactical or incomplete error in your input) ... and in engine.log I see 2016-08-10 11:59:43,780 WARN [org.ovirt.engine.core.bll.storage.pool.UpdateStoragePoolCommand] (default task-11) [6f626a79] Validation of action 'UpdateStoragePool' failed for user admin@internal-authz. Reasons: VAR__TYPE__STORAGE__POOL,VAR__ACTION__UPDATE,$ClustersList Default,ERROR_CANNOT_UPDATE_STORAGE_POOL_COMPATIBILITY_VERSION_BIGGER_THAN_CLUSTERS Instead in 4.0.2 it should allow me to do it so that I should see all running VMs, inlcluded the hosted engine VM, marked with configuration change needed symbol?

In this case for engine VM would it be sufficient to run something like thsi to commit its modifications? hosted-engine --set-maintenance --mode=global shutdown engine VM from os console hosted-engine --set-maintenance --mode=none verify engine VM is automatically started and without the flag about configuration changes needed?

On Wed, Aug 10, 2016 at 12:38 PM, Yedidyah Bar David <didi(a)redhat.com&gt; wrote: > > > > > > Ok, thanks for the update and clarifications. > > From an operational point of view, what will change? > > RIght now, in 4.0, if I select Default cluster and try to edit it and > > upgrade its Compatibility Version to 4.0, it seems I don't get any > > message > > in the gui, or at least I don't see it; it's something similar to when > > you > > have sort of syntactical or incomplete error in your input) ... > > and in engine.log I see > > 2016-08-10 11:59:43,780 WARN > > [org.ovirt.engine.core.bll.storage.pool.UpdateStoragePoolCommand] > > (default > > task-11) [6f626a79] Validation of action 'UpdateStoragePool' failed for > > user > > admin@internal-authz. Reasons: > > VAR__TYPE__STORAGE__POOL,VAR__ACTION__UPDATE,$ClustersList > > > > Default,ERROR_CANNOT_UPDATE_STORAGE_POOL_COMPATIBILITY_VERSION_BIGGER_THAN_CLUSTERS > > > > Instead in 4.0.2 it should allow me to do it so that I should see all > > running VMs, inlcluded the hosted engine VM, marked with configuration > > change needed symbol? > > That's what I understand too. > > > In this case for engine VM would it be sufficient to run something like > > thsi > > to commit its modifications? > > > > hosted-engine --set-maintenance --mode=global > > shutdown engine VM from os console > > hosted-engine --set-maintenance --mode=none > > verify engine VM is automatically started and without the flag about > > configuration changes needed? > > Looks ok to me, didn't try this myself. > > Best, > -- > Didi Hello, so I upgraded from 4.0 to 4.0.2 RC4 both engine and host. I successfully verified the cluster upgrade workflow; see here the dialogue: https://drive.google.com/file/d/0BwoPbcrMv8mvSVFFVzE5RTh5ZFU/view?usp=sha... Then the virtual machines pane appears like this: https://drive.google.com/file/d/0BwoPbcrMv8mvTXFnTDFVSllPaVk/view?usp=sha... Tried to start a previously powered off VM and it appears running without mark. Tried to shutdown and then powered on a VM with the configuration mark (c6 VM) and after correct start it doesn't have it The only doubt is related to hosted engine VM... it doesn't show any mark ... does this mean that actually I don't need to shutdown and power on the engine VM itself or is it an error that no configuration message icon is shown aside of engine VM?