Offloading Jobs by author to publish instances

[kamil_cieciersk]

Hi,

We have the case that it is needed to trigger service on one of the publishers which will connect to 3rd party app. We need to avoid scenarios that multiple publishers can call 3rd party so that it is considered to utilize Sling Job Offloading and create a topology of author and publishers (each instance is in different topology).

My question is if this is possible (and correct from a security perspective) to use the topology with authors and publishers.

AEM documentation suggests that offloading is only for authors:

for the offloading situation all of the instances involved are authoring instances.

On the other hand, it is written in Sling documentation:

If different clusters in the topology should represent different 'types of clusters' (eg a publish or an author cluster), then that is not explicitly handled by the discovery API. Instead, applications can define properties on each instance that model such cluster types or other aspects.

Does anybody try this scenario? Is it allowed, as instead of reverse replication only, we have also additional communication publish->author.

Kamil

[smacdonald2008]

I do not believe this is a recommended approach. The main way publisher communicates with author is reverse replication.

[kamil_cieciersk]

Actually, after checking the second time, job offloading seems to use reverse replication, too:

This setup requires the following three agents:

1. An outgoing agent on the author instance that replicates to the worker instance.
2. A reverse agent on the author instance that pulls from the outbox on the worker instance.
3. An outbox agent on the worker instance.

Is then any clear explanation why worker instance could not be a publish instance?