There is any way to make this operation faster?
I'm trying to restore my DB to the AWS DocumentDB, and probably it will take some weeks to finish... my overall data is less than 400MB.
dump is Gzipped
Asked
Active
Viewed 644 times
6
-
How are you restoring it? Which size are you using? – Chris Williams Jun 14 '20 at 07:37
-
@mokugo-devops sorry I didn't understand, which size are you talking about? – Nir Berko Jun 14 '20 at 07:40
-
Document db instance size – Chris Williams Jun 14 '20 at 07:43
-
@mokugo-devops `db.r5.large` and i'm restoring via this command: `mongorestore --uri="...." dump --gzip` – Nir Berko Jun 14 '20 at 07:48
-
@mokugo-devops using mongorestore from my MacBook to the DocumentDB – Nir Berko Jun 14 '20 at 07:53
-
2Could you test from using an EC2 instance to run the command? It might be quicker if it does not need to keep traversing the internet – Chris Williams Jun 14 '20 at 07:54
-
No problem, added this as an answer :) – Chris Williams Jun 14 '20 at 08:42
1 Answers
10
To resolve this it was suggested to run the command from an EC2 instance rather than a remote host.
This enabled a speedy import.
The likely reason is the number of network based operations across the internet rather than a local network resource which has shorter latency between each interaction.
Chris Williams
- 32,215
- 4
- 30
- 68
-
1You can speed up your mongodump by increasing the number of parallel connections using the --numParallelCollections option https://docs.aws.amazon.com/documentdb/latest/developerguide/backup_restore-dump_restore_import_export_data.html – Joseph Idziorek Jun 14 '20 at 13:18