AEM Replication queue | Community
Skip to main content
New Participant
September 2, 2022
Solved

AEM Replication queue

  • September 2, 2022
  • 5 replies
  • 2551 views

Dear Community,

 

Recently we have faced an issue with the replication agent for one of our publish instances, the queue was blocked making the authoring instance slow. we had disabled the agent and then re-enabled it when the pub instance is back. Now that the queue is gone and during the downtime, the replications didn't happen to this instance, how do we make sure that the pub instance is in sync now?

 

Thank you.

This post is no longer active and is closed to new replies. Need help? Start a new post to ask your question.
Best answer by joerghoh

I would clone an existing publish instance and replace the one where the content is incomplete.

5 replies

Chandra_Hire
New Participant
September 5, 2022

Hi @seeker ,

I could see below possible scenarios.

1. If resources which are supposed to be synced are minimal - then, identify such resources through query builder or logs and publish it manually. 

2. If resources are significant and unable to manage sync manually as suggested in step 1 - then, clone the other publisher (here need to handle the traffic carefully to the remaining publisher/s during cloning)

3. If 2 is difficult may be because of only 2 publishers or repository is huge or any other reason - sling distribution can help. Please check more on sling distribution - Sling Content Distribution In Aem (Part 1) | AEM ... - Adobe Experience League Community - 366602

 

Note: In order to avoid content sync issue especially when publisher needs to be taken off or agent needs to be disabled, I would suggest to have a dummy agent (on author pointing to fake publish url) enabled so that it will block the queue. Later point dummy agent to actual publisher (whenever it is up) so that blocked content will be published to the recently enabled publisher and we can avoid content sync issue.

 

arunpatidar
New Participant
September 4, 2022

I would suggest checking the replication status for the page/asset/XF during block time.

You can use any query like QueryBuilder etc and package the content from pub1 to pub2(blocked one).

 

It is also possible to publish again from Author, but that involves the risk of publishing further modifications that aren't ready.

Arun Patidar
Sachin_Arora_
New Participant
September 4, 2022

Ideally re-enabling the agent should still have Queue with all data. You can check logs of the time when queue was stuck. By default AEM generates log of failed resource. Please refer below log for reference. You can grep old logs referring this:

*ERROR* [sling-threadpool-3224e2e6-0696-45d3-a3ad-37e19a15df8b-(apache-sling-job-thread-pool)-1-com_day_cq_replication_job_publish(com/day/cq/replication/job/publish)] com.day.cq.replication.Agent.publish >> Path: /content/dam/asset.png

Also please check if you can find any information under /var/eventing/jobs

You can also try for getting information of resources that got replicated in that particular duration using a query and write script to publish then using Agent.

If its easy to clone the publish instance, then you can replace the outdated publish instance with the latest one. 

joerghoh
joerghohAccepted solution
Employee
September 4, 2022

I would clone an existing publish instance and replace the one where the content is incomplete.

Pradeep_Kumar_Srivastav
New Participant
September 3, 2022

You can manually migrate content and asset from one publisher to other through crx package manager to make sure both pubs are in sync. 

seekerAuthor
New Participant
September 3, 2022

is there a way to retrieve the missed activation and have them installed only?

Pradeep_Kumar_Srivastav
New Participant
September 4, 2022

Hi @seeker there is something called jcr checksum compare which can be used to see the difference but I would highly recommend to replace the existing content of publisher having issue to one which is correct.