Restore a sharded cluster from an AWS EBS backed volume (with the intent to create a new environment)

[Allan_Chase]

Existing env: RHEL 7 Mongo 6 Community Edition
Target env: RHEL 8 Mongo 6 Community Edition

Hello, I’m curious If I could build out a new clustered environment that is an exact mirror of my existing mongo environment by just copying one of my EBS /data1 and /data2 to one of the new mongo instances…will it sort of just…rebuild itself or do I have to copy all /data1 /data2 mounted drives on every EC2 instance and copy/mount that onto the new instances? It’s almost like a backup and recover some place else sort of thing (in my mind). I figured that before I put in a million, bajillion hours into trying it, some who’s done it might say…“dude, just copying one set of “/data1 /data2” isn’t going to cut it”. Hopefully I was able to articulate that question properly.

[Aasawari (Aasawari Sahasrabuddhe)]

Hi @Allan_Chase and welcome to MongoDB community forums!!

@Allan_Chase

I’m curious If I could build out a new clustered environment that is an exact mirror of my existing mongo environment by just copying one of my EBS /data1 and /data2 to one of the new mongo instances…will it sort of just…rebuild itself or do I have to copy all /data1 /data2 mounted drives on every EC2 instance and copy/mount that onto the new instances? It’s almost like a backup and recover some place else sort of thing (in my mind).

Based on the above statement, if you wish to restore a sharded cluster there could be more than one way to perform it.

1. You could use the file system method to backup the snapshot and then perform the restore for the sharded cluster in the target environment.
   Specifically, if your deployment depends on Amazon’s Elastic Block Storage (EBS) with RAID configured, you can follow the procedure to create the backup.

2. The other way to do it would be using the Cloud Manager for backup and restore which is a part of the MongoDB Enterprise version.

3. If your data is in Atlas and you wish to perform the restore, you can make use of Atlas live migrations to move the data between the source and destination environments.

Please let us know if you have any further questions regarding the same.

Warm Regards
Aasawari

[Allan_Chase]

Thank you Aasawari, I appreciate it. I think recommendation #1 is what I’m looking for. I don’t have the Enterprise version, so it sounds like I’ll have to get crafty in terms of addressing general backup and recovery. It sounds like I just need to shutdown our Prod, snapshot all data1/data2 mounted on every EC2 instance and share the snapshot/volumes with all the members of the new environment. In the end, I probably just have to really play with it to see what works best in our environment.