I have a SQL Server 2022 Enterprise Edition database called PCRData. It has 300 million rows spread across 260 tables. Each of the 260 tables are memory optimized tables and they collectively hold in memory, along with the non-clustered index, 34 gigabytes of memory. The max allocated RAM for SQL Server is 64 gigabytes. The filegroup contains one file that has an unlimited growth size.
I realize a 30 gig in memory database may seem like a lot, but it shouldn't be. The database appears to be configured properly, therefore, why is it that whenever SQL Server services are reset, the database goes into restore mode, then to suspect, then to rebuildable, then to "cant restore from backup". It fights me on every attempt at recovery, when I don't see a good reason for it to ever go into "Suspect" mode to begin with.
I need a solution to this or else I regrettably cannot use in-memory OLTP. I figured SQL Server 2022 would have this perfected by now.