Skip to Main Content
Spotfire Ideas Portal
Status Planning to implement
Categories data sources
Created by Guest
Created on May 31, 2023

Add full support for Azure SQL DB using Hyperscale

Provide support for Azure SQL DB [using Hyperscale] as a data source. Caching and data ship are required using native BCP in windows and linux.

  • Attach files
  • Guest
    Oct 19, 2023

    Additional aspects that must be considered while implementing the IDEA.

    • The Hyperscale HA will dynamically drop and create instances of the read-only replicas.

    • The issue with this methodology is that the TDV data source adapter and driver do not gracefully recognize when a replica goes away as the connection in TDV is still being held. What shows up in TDV is 08S01 which gets sent back to the client.

    • Making use of the "Microsoft SQL Server 2019" adapter and driver because there is no full support for Azure SQL DB and the HA capabilities.

    • What we would like to happen is for the adapter to gracefully handle the 08S01 connection is closed or connection is reset at the adapter level. It should close that connection and remove it from the connection pool so that other connections do not attempt to reuse it.

    • Even better would be for TDV to drop the connection out of the pool and retry with a new connection to Hyperscale so that Hyperscale can route to the new replica.

    • Currently, Hyperscale HA routes to a replica, and TDV will be holding a direct connection to that replica. Now Hyperscale destroys that replica and creates a new one but TDV does not know about the new one and still thinks the old one is viable. TDV does not remove that connection from the pool.

    • Once 08S01 errors start they keep happening because the connection is never released from the TDV connection pool.