ReportServer database procedures are continuously getting recreated in scale-out deployment

[daniel_doboseru]

Hi,

In a scale-out setup with two nodes, we started by having a standalone instance with May 2024 version, joined another one with Jan 2025 version, and started to have some transient issues (see related issue below). We followed the recommendation and brought both instances to Jan 2025 version by upgrading the first one too. They are now both at: Version1.22.9153.7886 (January 2025).

We are still encountering some transient errors with the message "An error occured within the report server database. This may be due to a connection failure, timeout or low disk condition within the database. (rsReportServerDatabaseError)" BUT we caught an instance where the detail message was: "Could not find stored procedure 'GetSnapshotPromotedInfo'".

We immediately asked the DBAs to check for this procedure, and it existed in the RS database, but its created_date was very recent (few minutes behind) and almost half an hour after we encountered the above message. Upon further checks it seems that this procedure is re-created regularly as created_date changes constantly.

We then checked another separate instance that we have for a few years now, and on that one the created_date is very old, even though the instance is heavily used (over 20k executions daily).

We believe that recreating these objects is actually not a normal behavior, and it might have something to do with the scale-out.

1. It's possible that the ReportServer database got somehow messed up because we had two versions of PBI RS connected to it?

2. What would be the best course of action to repair it without reinstall?

note: there were also other stored procedures that were initially not found which are most likely in the same situation - with this particular one we dug deeper to check it.

Thank you!

[daniel_doboseru · Accepted solution]

A feedback to the above problem: even though both instances were on the same version (Jan 2025) one of them was running in trial mode because of an issue with the key and the db objects had a different footprint thus causing the server to try to repair/re-sync them each time a request came from the other instance. After fixing the key issue and upgrading from trial to normal instance the problem disappeared instantly.

[lbendlin]

If you have a Pro license you can open a Pro ticket at https://admin.powerplatform.microsoft.com/newsupportticket/powerbi
Otherwise you can raise an issue at https://community.fabric.microsoft.com/t5/Issues/idb-p/Issues .