mongodump is very slow

Working with Data Developer Tools
,
1

Hello, I’m currently using mongodump for backing up my production MongoDB instance. However, the process is taking an unusually long time—about 34 hours to complete a backup of 750 GB of uncompressed data.

Here are the details of my environment:

MongoDB Version: 4.0.27
OS Version: Ubuntu 18.08
Server Type: Standalone server
RAM: 128 GB (with around 35 GB available)
Backup Method: mongodump with gzip compression

Could you please suggest ways to speed up the backup process?

Also, I would like to know the best backup strategy for a production environment of this size, considering factors like efficiency, reliability, and restore times.

2

If you use the Community Edition, the way to speed up the backup is to increase the number of parallel and use the parameter -j.

3

If spending some money is an option then Cloud Manager is a very good option that also provides point in time recovery over a configurable period. Incremental backup reduce backup time between full backups (depending on database churn).

Using a file system or volume snapshot is a very fast way to take a backup. These are also fast to restore as the indexes remain intact. This is my first choice after Ops Manager or Cloud Manager.

Other backup methods are mentioned on https://www.mongodb.com/docs/manual/core/backups but are of decreasing value with this size database.