What are the architecture features where CloudBase can work with abstraction layer such as SAN vendor(s) via CINDR driver (then SAN vendor agnostic) such that it can handle VM / PVC BCDR?
A LOT of vendors claim they can do failover, and they can. The control planes can handle workflow even with SAN as storage layer : VM -> Create Snapshot ->. SAN replicates SNAP to DR -> Hit big red button ---> Bring up VM by some service replicates volume / pv / snapshot ID over to DR cluster / site and can then bring up host / container.
BUT.. .and this is where everyone dies.... BCDR then comes in. All vendors fail the real hard / 1/2 half of this.
Now your in DR. Now call via CINDR to SAN and reverse replication in storage system, such that you do NOT have to do a bit full recopy, but can just get to last checkpoint that the production site has in sync.. then via snapshots and change block tracking the SAN does, catch it back to sync. (AKA return to data protecetion is months away .. recopy -> fail back to prod -> recopy again... ya.. stupid)
Without this your having to do a full bit recopy , which with the limited WAN and throughput of those syncs... it will Questions 1) Does CloudBase Coriolis work via abstraction layer such as CINDR to handle storage arrays and replications. Handle with storage vendor to get VM/ pvc to data consistent state, create snapshot. handle metadata of Snapshot disk ID handed up to hypervisor both local and DR. 2) Can CloudBase Coriolis work via CINDR to storage vendor to handle logic of, once site is back up, get reverse replication setup, validate back in sync, and then allow return back to production site (rinse repeat of resync via snap / data delta track and sync). So BCDR can be maitnained without 2 x full data copies).
