I am designing 2 datacenters that will be primary and data recover data centers. The roles of the sites will switch on a given internal to prove the process will work and for maintenance etc. It will also provide true DR if one site can no longer serve our customers.

The question is what if on the night of DST fall back (this year and prior that would be the last sunday in oct - the next 3 years it will be sometime in november) at 1:50 am the site (A) fails to the DR site(B) (1 hour earlier). Additionaly, there is an event that requires a PIT recovery after it failed to the DR site (B) say 1:49 am. (during the fall back hour).

KB 289472 says that 'SQLAgent creates a memory structure to support the xp_sqlagent_enum_jobs (This is what derives the next run date of sql agent tasks) action.' That memory structure is where Agent stores the calculated next date and time, and it can be lost during a cluster failover. So if this information is lost during cluster it will also be lost during a site failover.

So, is it, and if so how is it, possible to actually recover to the site (B) 1:49 am PIT?