[ovirt-users] Messed up upgrade to 4.2. Solution badly needed.

[Yedidyah Bar David]

On Fri, Nov 3, 2017 at 2:21 PM, Marcin Jessa <lists at yazzy.org> wrote:
> Hi guys.
>
> I have VM with hosted engine and a two node setup. When beta came out I was “OH! New shiny tech! Let’s try it!”. You all know the feeling.
> Unfortunately the upgrade process did not go as expected.
> I created a backup of my 4.1 installation and started to upgrade the VM. All went well. I then migrated all the running VMs to my second node put the cluster in global maintenance mode and updated one of the nodes.

Please provide more details.

You upgraded the engine vm to current 4.2 beta? yum update and engine-setup?

Then you moved a node to local maintenance, and updated it to 4.2
beta, using yum update?

Moving to global maintenance is for upgrading the engine vm, not a host.

Which documents did you follow?

> Then I run hosted-engine —upgrade-appliance on the node but it fails:

Where did you find instructions to use this tool?

It's meant only for migrations from 3.6 to 4.0, not general upgrades.

>
> [ ERROR ] Failed to execute stage 'Environment customization': exceptions must be old-style classes or derived from BaseException, not str
>
> The log file says:
> 2017-11-03 13:12:48,259+0100 DEBUG otopi.context context._executeMethod:143 method exception
> Traceback (most recent call last):
>   File "/usr/lib/python2.7/site-packages/otopi/context.py", line 133, in _executeMethod
>     method['method']()
>   File "/usr/share/ovirt-hosted-engine-setup/scripts/../plugins/gr-he-upgradeappliance/engine/misc.py", line 180, in _check_spm
>     'Unable to find this host in the engine, '
> TypeError: exceptions must be old-style classes or derived from BaseException, not str
> 2017-11-03 13:12:48,259+0100 ERROR otopi.context context._executeMethod:152 Failed to execute stage 'Environment customization': exceptions must be old-style classes or derived from BaseException, not str
> 2017-11-03 13:12:48,260+0100 DEBUG otopi.context context.dumpEnvironment:821 ENVIRONMENT DUMP - BEGIN
> 2017-11-03 13:12:48,260+0100 DEBUG otopi.context context.dumpEnvironment:831 ENV BASE/error=bool:'True'
> 2017-11-03 13:12:48,260+0100 DEBUG otopi.context context.dumpEnvironment:831 ENV BASE/exceptionInfo=list:'[(<type 'exceptions.TypeError'>, TypeError('exceptions must be old-style classes or derived from BaseException, not str',), <traceback object at 0x41d8128>)]'
> 2017-11-03 13:12:48,261+0100 DEBUG otopi.context context.dumpEnvironment:835 ENVIRONMENT DUMP - END

AFAICT a failure at this stage should be harmless - it's before doing
significant changes to the system.

Please share the log, just in case.

>
> I can login to the hosted engine but the nodes are down, the VMs are down and the storage is not connected.

Are you sure all was fine before you ran this tool?
I do not think your problems are a result of running the tool.

>
> Can you please advice what to do now?
> Is there a way to upgrade my setup?
> Should I disconnect my second node, which is also upgraded to 4.2 now and try to install hosted engine from scratch?
> Anything else I can do to save my VMs?

Not sure what your exact state is, or what is important to you and what isn't.

Are your VMs up?

Is it important to keep them up?

If not, you can reinstall everything from scratch and import the storage domain.

Otherwise, please provide more details - including relevant logs
(engine, vdsm, etc),
actions you took, etc.

Generally speaking, current 4.2 is not meant for production yet.
It's intended for people that want to see/test the new features etc.

Regards,
-- 
Didi