Local Cluster Connection Ids can not be unique between liferays nodes in a cluster

• Type: Bug
• Status: Closed
• Resolution: Fixed
• Affects Version/s: 7.2.X, Master
• Fix Version/s: 7.2.10 DXP FP5, 7.2.10.2 DXP SP2, 7.2.X, 7.3.10 DXP GA1, Master
• Component/s: Search Infrastructure
• Labels:
  • liferay-dxp-72-sp2
  • liferay-fixpack-dxp-5-7210
  • marketplace-portal-search-elasticsearch7-3.0.1-7210
• Branch Version/s:
  7.2.x
• Backported to Branch:
  Committed
• Fix Priority: 4
• Epic Link: Elasticsearch Cross Cluster Replication
• Sprint: Search|S03 Sprint 13| 3.16-4.6
• Git Pull Request: https://github.com/BryanEngler/liferay-portal/pull/495

1) set up 2 liferay servers in a cluster, one running on port 8080 and the other on 8081

in portal-ext.props set
cluster.link.enabled=true and point both to the same database

optionally set the config to share the same document repo to prevent any file not found errors

2) set up 3 separate elastic clusters. configure one cluster as a "leader" and set up the other 2 to replicate from it

3) start liferay and add 3 elasticsearch connections - "remote" (pointing to leader), "ccr1" (pointing to replica), and "ccr2" (pointing to other replica)

4) set opertation mode to REMOTE and use the "remote" as the connection Id

5) go to Cross cluster replication, enable it and add 2 configs (use + button for second config)
localhost:8080=ccr1
localhost:8081=ccr2

6) go to the "connections" tab in Search admin. verify the "write" label is next to the remote connection for both the 8080 and 8081 server. on the 8080 server, verify the "read" label is next to the ccr1 connection, and on the localhost:8081 server, verify the "read" label is next to the ccr2 connection

7) verify that added content from either liferay server is searchable by both liferay servers.

8) verify that server 8080 can still search from ccr1 when the ccr2 ES server is shut down, and that 8081 can still search with ccr1 down and ccr2 up

9) verify full reindex works