NPM 2023.4.x Upgrade feedback

Upgrade process was fairly smooth. We had an issue with database maintenance but not a major one

We have upgraded 3 of our environments to 2023.4.2 now and all went smoothly with no obvious issues and certainly nothing that we've spotted or had reported to us. two small, single server deployments and one multi server, multi APEs and HA.

I will add if you are on 2023.2.x or .3.x then make the jump. It fixes loads of bugs and made our systems stable again.

Hi stuartd 

I am currently on 2020.2.6 and planning to upgrade to 2023.4.2, I know i can directly upgrade to the 2023.4.2

I was going through a few threads:

1. Backup any Unmanage utility tasks

2. Make sure there are no custom queries used in syslogs/traps rules

3. Is there anything else that you can think of that is deprecated or needs to be recreated ?

Direct from 2020 I don't think you need any specific steps outside of the usual 'backup your database' - but your extras are always a good shout to check or backup.

We had no issues with it

I just made our third attempt to get off of 2020.2.6, this time to 2023.4.2.  Again, received a failure during the Configuration Wizard stage that was yet again different from previous attempts to previous versions:

"Database configuration failed:  Error while executing script- Column already has a DEFAULT bound to it.  Could not create constraint or index.  See previous errors."

I'm finally opening a case as we need to get this thing upgraded as I can't find anything on this error.  I don't know why these recent upgrades are so problematic for us, we aren't doing anything particularly custom.  Also annoying since we completely rebuilt our cluster from scratch a few years ago.  I've easily got 120 hours into trying to get past 2020.2.6 and it's become annoying.

In the CW logs they usually post exactly the SQL they are trying to run at each step, so if I run into something weird like that I usually just dig into the SQL and reverse engineer the change that it's attempting to make and see what hoops I need to clear to make it happen.  In the past I've had to do various fixes for customers that didn't seem like they would have been needed, but things like dropping existing tables just for the CW to recreate them, or removing indexes, other bits.  Can't say why some environments run into things like that when most don't but it seems to happen maybe one out of 20-30 upgrades that I've performed for clients.

My issue is that I'm a network guy, not a SQL whiz.  I know my way around servers well enough that I'm the defacto "server guy" on our team, but issues like this are precisely why we pay $$$ for support.

I was able to get past this issue by deleting one of the constraints in SQL.  It sounds like support was super familiar with this issue as I didn't even provide logs or diagnostics.  They were able to give me a working procedure just from the error message.  We are finally off 2020.2.6, so I'm happy.

Starting with 2023.2, new alert actions that require executing an external program or VB script require an approval by a system administrator.

I had some alerts that used scripts to create trouble tickets in our ticketing system.  After the upgrade the scripts stopped working.

Approve alert actions executing a script (solarwinds.com)

documentation.solarwinds.com/.../core-approve-execute-alert-action.htm

still issues with the database maintenance on upgrade ? DB auth issues ?