Let say i have 2 Open-E storage servers - master and slave, they will be active-passive Synchronous replicated HA enabled. The master will have Raid card with SSD caching. The slave will have software raid.
The master storage raid card SSD cache will cache the most frequent used files. It will works as READ and Write cache.
What happen with write cache when the main storage goes offline and the slave takeover the master's function and then after few hours master comes back online.
What will happen with cache data if it didn't got time to flush it to main storage ? It will get corrupted ?
The point is: Is it stupid to use write cache with Synchronous replication in order to bypass data corruption or data mismatch in both nodes in case if something goes wrong with master and its goes offline for 2 days? For sure, the new data will be written in slave in those 2 days.
I would like to hear your thoughts about that stuff. I hope you guys understand i'm trying to say.
SSD cache and failover
Reply With Quote