aem author to author content replication for sync

[sayabannatole]

Hi,

We are looking at available solutions to sync two author instances and one of the option we thought of using author to author replication.

Can we use this option to sync content between two instance? will this option have any implications on the performance or security?

any inputs would be great.

Thanks 

[shelly-goel · Accepted solution]

@sayabannatole  I would suggest to use Grabbit for migrating/ syncing content between two instances. It's much faster and robust.

Replication would have performance issues with so many nodes to be replicated at the same time and would slow down the instances.

https://github.com/TWCable/grabbit

[Rohan_Garg]

Hi @shelly-goel,

Query 1 - Did you have any issues in using Grabbit with AEM 6.5 ?

As per Grabbit's github repo latest version is v7.1.5 released 2 years ago and support is mentioned till AEM version 6.3.

https://github.com/TWCable/grabbit/blob/master/docs/AEMSupport.adoc

[Rohan_Garg]

Hi @shelly-goel,

Query 1 - Did you have any issues in using Grabbit with AEM 6.5 ?

As per Grabbit's github repo latest version is v7.1.5 released 2 years ago and support is mentioned till AEM version 6.3.

https://github.com/TWCable/grabbit/blob/master/docs/AEMSupport.adoc

[shelly-goel]

I did use it on AEM6.3 but I would suggest you can give it a try on AEM6.5

[sayabannatole]

Hi @shelly-goel  Thanks for your response i will definitely have a look at using Grabbit. My concern about using author to replication is still workable option right , barring the performance issues which can be accommodated with enough resources on both instances? 

[shelly-goel]

Conceptually it can be done though performance would not be good with tree replication (can be better with more compute resources). You can rather create a content package, download and install on another instance.

[Nirmal_Jose]

There is no problem in using author replication agent, but it is an additional overhead to production author when each page is published. You can use the below flow chart to select the right method for content synchronisation.

Pros and cons of each method is depicted as below. 

Hope this helps to select the right approach for you.

[sayabannatole]

Hi @Nirmal_Jose 

Thanks for the clarification, couple of things i have observed for two of the options i.e in case of backup/restore if there's a segment error in source then it will get carried over as part of sync to target..in case of vlt-rcp the sync fails occasionally due to unknown reasons and needs to remove the mapped path and resync everything from the beginning

[Nirmal_Jose]

Yes, indeed. Whatever we do, we need to create repeatable infrastructure as much as possible. Backup and restore can be as simple as a crx-quickstart folder zipped and copied over OR if a cloud infrastructure in AWS, use a snapshot of the ec2 instance. Using replication, there are various other scenarios like a) JCR properties/status show related to replication will be shown incorrect are not relevant. b) Unless published, we won't see the content