Environment Display is messed up after upgrading Octopus 2.6.4 to 3.2.24

Hi

When an environment contains at least 2 target deployments and there is 1 machine (A) which has another machine (B), which appears on the right hand side to machine A, and machine A has a name which is long enough - lets say called: “CantDisplayThisLongNameInOctopus3.x” the display of the environment is messed up- please have a look at the attached screen shots

The problem is reproducible also in version 3.2.19

Thanks
Nir

Hi Nir,

Thanks for the heads up. It seems the Machine name does break properly if you include special characters like hypens or even spaces (as seen in your first screenshot), but not if its a very long name with no kind of breaks. Given that this is the first time we hear about this issue, it seems like machine names without special characters or spaces are not that common.

We’re gonna let this one sit until our backlog gets small enough so we can tackle purely visual things such as this one.

Warm regards,
Dalmiro.

Hi Dalmiro,
Thanks for your answer.
This might cause a big problem for us as part of our migration.
We have a few dozens of environments and in each one there are about ~35 machines, which are currently displayed properly with long names (machine qualified domain name).
Can you maybe provide a workaround (to rename these machine names on the raven (2.6)) before the import?
If we go on with the upgrade as is and you aren’t planning to fix this sometime soon this might make our lives very hard…

Thanks
Nir

Hi
Just wanted to update you that I am running the following update when I finish the migration to work around:

Update [Octopus].[dbo].[Machine]
SET Name = REPLACE(Name, ‘.’, ‘-’)

Thanks
Nir

From: Nir Weinberg (ITaaS)
Sent: Thursday, 10 March 2016 13:30
To: 'Dalmiro Grañas’
Subject: RE: Environment Display is messed up after upgrading Octopus 2.6.4 to 3.2.24 [Problems #44846]

Hi Dalmiro,
Thanks for your answer.
This might cause a big problem for us as part of our migration.
We have a few dozens of environments and in each one there are about ~35 machines, which are currently displayed properly with long names (machine qualified domain name).
Can you maybe provide a workaround (to rename these machine names on the raven (2.6)) before the import?
If we go on with the upgrade as is and you aren’t planning to fix this sometime soon this might make our lives very hard…

Thanks
Nir

Hi Nir,

I was going to recommend you exactly that: Try doing the renaming when you move to 3.X from SQL instead of doing it from Raven.

I’ve created a github issue to get this logic reviewed. Do keep in mind though that we work on issues based on priority, and it’ll take a while for this one to climb its way to the top.

Best regards,
Dalmiro

Hi,
Totally understood, but I would think that machine fqdn is pretty common.
As I was able to work around this issue it’s not very important any more from my POV, but I just thought it’s good to share it with you and maybe other users that would encounter this

Thanks
Nir

Just an update on this:
It turned out as not such a great idea.
I was doing the migration yesterday struggling with hydra and other nasty issues, during that stage I’ve noticed that in most of my environments I have double the machines I need to have. The final import didn’t overwritten the ones with ‘-’ but added the ones with ‘.’ on top of them
I had to run a delete sql query and only then run the sql above