Too bad they did it wrong, RecoverPoint is truly an awesome product and is making my life significantly easier. Nice idea, EMC, really nice idea that will hopefully be fixed in SP 2.
RecoverPoint is a fabric based data replication technology that can either replication synchronously, locally or asynchronously remotely. The tool enable a point in time recovery at the replication target that appears to only be rivaled by SRDF. Word on the EMC street is that SRDF code is being replaced/augmented by RecoverPoint in the V-Max series SAN. I’ve been told that the compression algorithm is superior.
Back to the point, Snapshot Consolidation is a feature enabled in version 3.1 that allows the consolidation of snapshot at user defined intervals. I see this as a huge benefit in Journal capacity needs. My instructions have been to provide a two weeks history of snapshots for recovery. Unfortunately, I don’t agree with this and see it as a waste of my valuable disk space. In my case, to attempt to save some disk, I’ve configured RecoverPoint to consolidate all snapshots older than 4 days into one daily snap and maintain two weeks history (almost the defaults). I also have configured daily and monthly snaps. I personally am not very interested in those as I try not allocate the SAN journal space to plan for retention of snaps that long, the business importance of a per write transaction diminishes significantly very quickly. My thoughts are that tapes exists to service those needs. My data is primarily a BCDR tool with the neat feature of being accessible for TEST/DEV data refreshes from PROD.
Back to the point, again. The snapshot consolidation feature will wrap up all writes to a location on disk to it’s end state and remove all previous snapshots between consolidation intervals. Per EMC:
When snapshots are consolidated, all writes to a single disk location are
saved in their end-state at one specific point-in-time. Therefore, all writes to
the same disk location between the start-time and end-time of the
consolidation are lost.
There is another feature which, up until now, I have not fully understood. I, apparently, have not found the documentation the follow source has. The feature is “Consolidate Snapshots Between Bookmarks” and is documented by Joe Kelley, who’s blog I’ve just discovered. Read for more insight on why this might not be desirable.
All of this sounds really cool and I was pumped to get it implemented despite the fact that it seems to cause a full volume sweep and eliminates journals. Whooosh, there goes 8 TB across the WAN! All thanks to RecoverPoint’s compression it really hasn’t pissed off my network admins yet. After implementing I received error reports of journal write failues and volume sweeps almost daily. After opening a case with EMC I was told this is a known bug with a work around of “disabling snapshot consolidation”. Thanks guys, let’s try this again.
SP 2 should be out soon, if it’s not already. Hopefully this will be resolved! I want my disk space back.