Looking for understanding - strange behavior after moving SCSM operational database
Last night we migrated the SCSM operational database from a SQL2012 instance to SQL2016.
The Cireson Service Portal broke because in our install, we had used a server name in the connection string instead of an alias. Thus, Cireson was no longer looking in the right place for the ServiceManager database.
I ran the Service Portal setup on our cache builder server, pointed it to the new location, and all was well. (This was a newer version of the portal, as I couldn't find the previous versions on the downloads page.)
In addition to the cache builder server, we have 6 other servers running the portal. I upgraded two of them. So far, so good. But within 15 minutes, we were broke. The portal was again trying to connect to the old db location.
I reran setup on the cache builder and we were fixed-- for about 15 minutes. This same thing repeated twice more, with it all breaking, then being fixed by a reinstall on the cache builder server.
We never determined what kept setting the database location back. However, when I completed the upgrade on all the other servers, the behavior ceased, and we've been good ever since.
Could it be that the non-upgraded servers were passing an old connection string back to the cache builder server (and thus breaking everyone)?