best practice for backup / restore

Unfortunately at the moment this is one of our weak spots. I've already engaged our engineering team and we are working on a definitive plan in place to offer backups and restore but I don't have a firm date for availability yet. We want to be able to supply this feature in an easy-to-use, robust and scalable mechanism.

We are planning on releasing a backup/restore feature in 1.6.1 which is currently scheduled for the end of October. The feature will allow for the pausing of persistence for a brief period of time while it makes a copy of the data on disk. All the functionality for this is there already, just needs to be more readily exposed to the user. This will allow for very quick backups to be taken, but (without extensive work) will limit you to restoring the data into an identically configured cluster.

I'll keep sending updates as I get more information and the feature develops.

Thanks a bunch, glad to hear you're having a good experience with Membase.

Perry

Thanks Perry - I have a couple follow up questions:

1) How do we turn off persistence temporarily then re-enable it?
2) When we do a backup of the disk, do we need to back up all disks on all nodes (since the data's distributed)?

Shawn

Just updated my original post to more align with our product management plans.

To answer your questions:
1 - I have to double-check, but I don't believe it's something that is controllable by the user today.
2 - Yes, you will need to backup the data files on all the nodes.

Thanks again.

Perry

Great, thanks for the update Perry. Will upgrade from 1.6.0 GA to 1.6.1 require a data/config wipe?

Shawn

Yes, you will be able to upgrade from GA to 1.6.1. It will be a "rolling" upgrade in the sense that you will have to rebalance one node out of the cluster, upgrade it and then rebalance it back in. Rinse and repeat.

Perry