Originally published in November 2020, at ObjectRocket.com/blog
At Rackspace, the ObjectRocket team can help you manage your MongoDB® databases. We offer disaster recovery, replication, fault tolerance, and high availability for your MongoDB database.
ObjectRocket currently offers Sharded and Replica MongoDB instance options, but behind the scenes, we always use three-member replica sets for data redundancy and fault tolerance. Sharded instances consist of dedicated mongos servers, config servers, and the shards themselves, which are each a three-member replica set.
However, disaster recovery is a different matter. ObjectRocket offers a disaster recovery scheme where your Sharded and Replica MongoDB instances receive mirrored member replica sets in another data center at an additional cost you can use in case of a disaster event.
All ObjectRocket managed MongoDB instances come with high availability, replication (data redundancy), and fault tolerance as a package.
With high availability (HA), you can access your data at all times through various technologies. However, a key component to almost any HA solution is a replica of your data. This is where data redundancy comes into play. You only see one database, but behind the scenes, there are two or more exact copies (replicas) of that data, which helps with the shift from one replica set node to another in the case of failure.
Fault tolerance is the property that enables a system to continue operating properly in the event of failure of a component within the same fault zone. To ensure business continuity, we use different hosts or servers for each of your shard’s replica-set members, so the others are operational if one goes down the others.
MongoDB replication is the process of synchronizing data across multiple servers at an application level. Replication provides redundancy and increases data availability with multiple copies of data on different database servers. Replication protects a database from the loss of a single server and allows you to recover from hardware failure and service interruptions. With additional copies of the data, you can dedicate one or more to disaster recovery, reporting, or backup.
Disaster Recovery (DR) is an area of planning that protects an organization from the effects of significant negative events like natural calamities, power failure that can cause data center-level outage. The goal of DR is for a business to continue as close to normal as possible.
Let’s assume a scenario where you have high sales throughout the year, especially on Cyber Monday and across the holiday seasons.
As a United States-based company, you want to store your data within the United States, so let’s use our DFW data center to store your data and the IAD datacenter to keep a backup.
We achieve HA by mirroring your data to two nodes that are acting as secondary nodes. If the primary fails for any reason, a secondary becomes the new primary node but still has another backup. In the meantime, our support team recovers the node that failed and syncs it with the new primary.
MongoDB can scale out horizontally through single large Replica sets by using one primary and two secondary nodes with heartbeat communication for up (or down) states and replication to the secondaries occurring through the oplog.
We achieve fault tolerance by having each host on different hardware within the same datacenter. That way, if any hardware fails, your MongoDB remains operational.
Generally, a shard could be a single MongoDB instance. However, as previously mentioned, replica sets on ObjectRocket include three MongoDB instances, as shown in the following image:
In case of maintenance, downtime, or failure, the system can switch the primary with any of the secondary nodes and continue without issue, thus providing HA.
Data is replicated from the PRIMARY node to the SECONDARY nodes through the oplog, thus achieving data redundancy.
As for Fault Tolerance, let’s imagine the unlikely scenario that the server that hosts one of the Replica set members breaks down. Whether a primary or secondary node failed, your MongoDB instance continues to run in a two-node replica set until the support team can remediate the issue.
Our scheme mirrors your current instance (regardless of the number of shards you have added to your instance) in one of the closest data centers to the instance. We also have a DR plan set up in IAD for you. WE can connect the IAD replica-set members to the primary node in DFW3, with your data replicated synchronously across regions. This ensures your data is up to date, as illustrated in the following example:
Use the Feedback tab to make any comments or ask questions. You can also click Let’s Talk to start the conversation.