Many times people run single instance databases without replication with various RDMS. Is something different with mongo that makes it highly or somewhat unreliable without replication?
Hi Adam,
For most applications, data is a critical commodity. Storing data in one place is a risky proposition, so you need to have a strategy or a plan in place to ensure that you can recover from a failure as quickly as possible.
Achieving High Availability through replication is not a concept that’s unique to MongoDB, or NoSQL databases. Some RDBMS also supports data replication feature for high availability.
High availability is very valuable in sharded cluster deployment, without it the cluster may be missing a large part of data while servers are in repair.
I rarely run replication on my RDMS: just a good backup strategy,raid10, and a watchdog script that will restart the database if the process crashes… is this not enough for mongo?
MongoDB replication is relatively easy to setup to provide fault tolerance. In your example, there are cases that may need extra care, such as the watchdog script server could have died, or there’s a corruption of the database file after an unclean shutdown that requires manual intervention, etc.
For more information please check out :
New to this group and new to Mongo.
I would recommend to enrol in a free online course at MongoDB University to learn more about MongoDB.
Regards,
Wan.