Setting up Replication Agent when publish instances share same MongoB | Community
Skip to main content
K
Katrien1
Level 3
October 16, 2015
Solved

Setting up Replication Agent when publish instances share same MongoB

  • October 16, 2015
  • 11 replies
  • 2445 views

We will be setting up one author and two publishers. One MongoDB for author, One MongoDB for both publisers with a replication set. In other words, Pub1 and Pub2 will be reading from the same DB. This is per recommended deployment from Adobe, and confirmed with Adobe consultant,

Question is around replication agent. We could have configure one agent that updates the publish environment via Pub1. But what if Pub1 is down. The consultant told us that there is a way to configure your replication agent with two URLs. THe first one will be taken, if that one doesn't repond the second one will be used to replicate.

However, I cannot find any details on how to do this. Anyone here done this before? I know I could also use a loadbalancer. But I'd like to hear solutions where you configure failover via the replication agent.

Thanks.

This post is no longer active and is closed to new replies.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

When you configure the replication agent, you can add multiple target URLs by adding them into the single "target address" field, separated by comma (","). In case the first target fails, the second (third, fourth, ...) is tried.

Jörg

11 replies

joerghoh
Adobe Employee
joerghohAdobe EmployeeAccepted solution
Adobe Employee
October 16, 2015

When you configure the replication agent, you can add multiple target URLs by adding them into the single "target address" field, separated by comma (","). In case the first target fails, the second (third, fourth, ...) is tried.

Jörg

K
Katrien1Author
Level 3
October 16, 2015

Hi, sorry to get back to this, but the option you suggested doesn't seem to work for me. I've got it configured as follows right now:

http://localhost:4503/bin/receive?sling:authRequestLogin=1,http://localhost:4504/bin/receive?sling:authRequestLogin=1

When I bring down the pub instance on 4503, the queue gets blocked and doesn't process the request to 4504.

Did I put the comma at the wrong location?

Cheers,

Katrien

K
Katrien1Author
Level 3
October 16, 2015

Thanks Jorg. Is there any documentation online that support this?

K
Katrien1Author
Level 3
October 16, 2015

No one?

I've been told by an Adobe consultant that there is a way to configure it that way. The configuration proposed to me doesn't work. I tried Daycare, but they too have no clue. It's pretty vital, in my opinion, that you can configure a replication agent the way I desire, when using master databases.

Sham_HC
Sham_HC
Level 10
October 16, 2015

jcbsktrn1 wrote...

Thanks Jorg. Is there any documentation online that support this?

 

Filled document request to update. The internal reference is DOC-4650

Sham_HC
Sham_HC
Level 10
October 16, 2015

Katrien wrote...

No one?

I've been told by an Adobe consultant that there is a way to configure it that way. The configuration proposed to me doesn't work. I tried Daycare, but they too have no clue. It's pretty vital, in my opinion, that you can configure a replication agent the way I desire, when using master databases.

 

It is supported feature to have multiple target URLS for replication  and should work. Otherwise would consider as bug.  Could you please send me offline the daycare ticket number? 

joerghoh
Adobe Employee
joerghohAdobe Employee
Adobe Employee
October 16, 2015

Another option would be to add a loadbalancer between author and publish, which handles the replication. The loadbalancer will always route the request to a running publish instance.

Jörg

K
Katrien1Author
Level 3
October 16, 2015

Hi Jorg, that would indeed be a solution and we will use that solution if it cannot be configured any other way. However, both the consultant and you (see above) have told me that you can configure using multiple targets, are you now saying that it's not possible? Adding another loadbalancer just complicates the architecture and I want to keep it simple :)

K
Katrien1Author
Level 3
October 16, 2015

Hey thanks Sham, I will send you the daycare ticket as a message.

Sham_HC
Sham_HC
Level 10
October 16, 2015

Katrien wrote...

Hey thanks Sham, I will send you the daycare ticket as a message.

 

Your settings are correct & I have demonstrated in video at [1],  Probably you might not waited for a retry timeout. I have put internal notes on ticket & case owner will take care of rest. 

[1]    http://aemfaq.blogspot.com/2014/07/replication-agent-with-multiple-target.html

Show more replies
Terms & ConditionsAccessibility statement

Loading footer...