Our company is getting ready to migrate from SAGE 100 2023 to SAGE 100 2026. Our current configuration is using a separate server for both the application server and one for the SQL server. We will be migrating to SAGE 100 2026 using the same configuration on new servers, application server and SQL server

After setting up the new SAGE 100 2026 server and getting it connected to a new SQL server I started the Parallel Migration Wizard to bring over a current copy of our production data do the users could test it out for a month or so before we upgrade. The wizard seemed to do fine and brought over all our companies and configurations.

Since these same servers will end up being our production servers once I pull a final copy of the data over again I set up both the App server and SQL with backups using Veeam. Veeam saw each server and start backing them up. After a day or two I went back to check the backups and make sure they were backing up correctly and able to restore data and databases if needed. That is when I saw that Veeam was skipping the backup on the SQL databases that were created during the data migration.

I worked with Veeam support for few days and then found the issue. When the SQL databases are created using the Parallel Migration Wizard the new database files are created in the correct location set by the installation of SQL but when the SQL databases are added to the SQL server and mounted they are connected to using UNC instead of the local drive location...

\\sqlserver\d$\sqldata\mas_data.mdb instead of D:\sqldata\mas_data.mdb

In this example you can see that the file can be found using this method but has 1 minor setback and 1 major setback.

The minor setback is that when the SQL server is connecting to these files it is using the UNC path and adds overhead to the connection just by doing so. This may not cause much of a performance hit on a small implementation but in a larger one it will add up and slow down everything.

The major setback is that when you use Veeam and tell it to back up the SQL server. It will include any databases that are hosted on the local SQL server. If the databases are referenced by UNC name they will be skipped and not backed up at all. The UNC name tells SQL that the database file is located somewhere on the network and can be ignored since it is probably not on the local server. If I had not checked my backups and seen that it was not backing up any of these databases. I am concerned that others have been migrated and since the server comes up and works just fine that everyone will assume that the backups are fine as well. Someone will find themselves with a broken database and a useless backup at some point.

In my instance I was able to take each of the databases offline and change the mount point to be the local drive letter instead of using the UNC name. This allowed my backup job to see the databases as local and start backing them up.

This will be an issue at some point for users who don't realize they aren't getting a backup of their data!!!

Please alter the behavior of your Parallel Migration Wizard so that it creates the databases and mounts them using the correct local path for the SQL server.

by: Roger Z. | about a day ago | General Enhancements