I use Gluster snapshots before patching the engine.
Usually the flow is:
1. Global Maintenance
2. Power off the HostedEngine VM
3. Gluster snapshot
4. Power on HE
5. Upgrade of the setup packages
6. Run the hosted-engine upgrade script
7. Patch the OS of HE
8. Reboot
9. Remove Global Maintenance
If you need to revert a snapshot, you need to stop the gluster volume , so you need to
follow the rule and keep the engine on a separate gluster volume.
Best Regards,
Strahil Nikolov
На 10 юни 2020 г. 13:21:08 GMT+03:00, Yedidyah Bar David <didi(a)redhat.com> написа:
>On Wed, Jun 10, 2020 at 1:05 PM <jrbdeguzman05(a)gmail.com> wrote:
>>
>> Hi Guys!
>>
>> How do you usually backup your hosted engine prior minor updates?
>>
>> I’ve tried to perform a minor update from 4.3.8 to 4.3.9 recently.
>I’ve used engine-backup before updating and for restore. However, the
>problem I’m encountering now is that, it looks like the hosted-engine
>settings are only being restored. The version is still 4.3.9 when I’m
>trying to revert it to 4.3.8.
>>
>> Am I missing some steps aside from engine-backup —restore? I saw on
>old posts that snapshots of hosted engine is not a recommended step
>before upgrade/update.
>
>engine-backup --mode=restore only restores the data. If you want 4.3.8
>packages, you have to install them. For rolling back packages on an
>existing machine, you can usually find the relevant yum transactions
>and 'yum history undo $ID' on them.
>
>If you installed and set up e.g. 4.3.8, and take a backup, then try to
>restore this backup on 4.3.9, the next engine-setup you run (which is
>a required step after restore) will upgrade your database
>schema/content to 4.3.9.
>
>Hope this helps,