The "Correct" way to back up on AWS with Membase in EBS?
I currently run two Membase nodes in Amazon EC2. The actual data store for these nodes is located on a seperate EBS drive due to their redundant storage and persistence after a crash. Also, the actual root drive is located on a smaller EBS drive.
After spending 30-40 hours trying to get Membase to back up using the provided mbbackup command, including posting here as well as in the sqlite mailing list, I have come to the conclusion that the mbbackup command is not the correct tool to use on the cloud in this situation. I figured this out in part by looking at the RightScale script for backing on the cloud (they have a free tier you can use even if they are not managing your cloud). Rightscale obviously knows what they are doing and works closely with CouchBase. In their script, they clearly just do some sort of an EBS copy, and don't bother with any of the sqlite3 commands.
My questions:
- Is simply taking an EBS snapshot of your drive, assuming the write queue is short and I copy Config.dat to the same EBS, enough to back up the needed data for restore?
- If not, can someone outline what would need to be done to do this correctly?
David, EBS snapshots are in fact a valid way of backing up. I'm sorry you've had so much trouble with sqlite.
Using mbbackup "should" work in any environment, though clearly it's not working too well for you. mbbackup gives you the ability to take the backups off-node, whereas an EBS snapshot would require more effort on your part.
Either one is really fine depending on your needs and environment.
Perry
Forum support is great for free but sometimes you need a guaranteed response time and dedicated resources for your questions or issues.
Consider purchasing enterprise-level support from Couchbase: http://www.couchbase.com/products-and-services/overview
Call or email "sales -at- couchbase-dot- com" today!