i should able copy renditions of asset worker instance master instance , delete asset inthe worker instance using dam update asset offloading workflow
in opinion not practice update update asset workflow on worker instance -
- this whole offloading based on sling discovery , eventing mechanism. requires offloaded asset sent (read reverse replication) leader instance
- adding step within update asset workflow may cause issues reverse replication of asset.
you have build independent of offloading process achieve deletion. there multiple ways -
one possible way -
- have jms based implementation monitor reverse replication
- if reverse replication successful either delete asset or mark asset deletion (highly recommended)
- if following approach of marking asset deletion, setup cleanup task run of worker instance (scheduled convenient time). cleanup task identifies assets marked deletion , processes them.
imho marking asset deletion better approach more performant , efficient. assets processed @ once during off-peak time.
there other ways require lot of custom code written.
updates -
tapping reverse replication -
you need details of working of reverse replication.
- content reverse replicated pushed outbox
- if @
/etc/replication/agents.publish/outbox/jcr:contenton local instance, propertytransporturidefault -repo://var/replication/outboxi.e. content reverse replicated pushed '/var/replication/outbox' - now @
/libs/cq/replication/components/revagent/revagent.jsp, logic works on receiving instance.
going through above give deeper understanding of how reverse replication working.
now have 2 options implement want -
- to check replication status, tap replication queue code in
/libs/cq/replication/components/revagent/revagent.jspdoing. code executes on author instance content reverse replicated, in case leader instance. have works around code make work on worker instance. more specific on implementation code update lineagent agent = agentmgr.getagents().get(id);id outbox agent id. - have event listener monitor outbox. check payload comes replication , use functionality.
what have mentioned crude approach doesn't cover failover/recovery use-cases, i.e. how handle deletion if replication queue blocked reason , images have not been pushed leader.
Comments
Post a Comment