Let’s explore how MongoDB replica set and Cloud Block Store in AWS delivers value to your business.
Do you need to build MongoDB replica set in AWS? Cloud Block Store will speed it up for you. Lock the file system on MongoDB node, copy the Cloud Block Store volume and connect the new volume to the replica set node. Start MongoDB and join the node to replica set. That is it! And do not forget to unlock the file system.
Do you need to grow your MongoDB replica set? No problem. Lock the file system on one of the secondaries, copy the Cloud Block Store volume, connect it to the new node. No performance impact, no database stopping, no downtime. And again, do not forget to unlock the file system.
Database volumes growing? Running short on disk space? No problem, we can trim them. During our testing we have observed that Cloud Block Store and FlashArray data reduction in MongoDB replica set is improved by overwriting existing volumes with a snapshot. So again, lock the file system on one of the secondaries, snapshot the database volume. Unlock the file system and lock it on another secondary. Unmount and overwrite the volume from snapshot. Repeat for the remaining replica set members. You will notice that Cloud Block Store data reduction is better.
Did one of the nodes fail? Need to recover as fast as possible? No problem. Lock the file system on one of the secondaries, copy the Cloud Block Store volume, connect it to the recovered node. The mongod process will recognize data on the new volume and change the node status to “Secondary”. No performance impact, no database stopping, no lengthy data copy over the network.
Cloud Block Store with excellent data reduction and efficient snapshots offers many advantages for MongoDB replica set deployments. Data reduction and rich set of data services are always included with Cloud Block Store. Now you can run your mission-critical applications in the cloud with 11 9’s durability from Amazon S3 and high IO1 performance with space savings from Cloud Block Store.
Need MongoDB information? Pure Storage blogs are here and, of course, MongoDB documentation: