Revert fail-fast disconnect strategy for _resolve/cluster (#124241) (#124368)

* Revert fail-fast disconnect strategy for `_resolve/cluster`

This was the first solution for the long wait times that user could
potentially see when invoking this API. However, we reverted this change
in favour of the new timeout parameter. Unfortunately, the change in
disconnect strategy targeted more broader versions than the timeout
parameter PR (which contained the revert). This PR fixes this discrepancy.

* Update docs/changelog/124241.yaml

Author: pawankartik-elastic

docs/changelog/124241.yaml

@@ -0,0 +1,5 @@
+pr: 124241
+summary: Revert fail-fast disconnect strategy for `_resolve/cluster`
+area: Search
+type: bug
+issues: []

server/src/main/java/org/elasticsearch/action/admin/indices/resolve/TransportResolveClusterAction.java

@@ -140,7 +140,7 @@ protected void doExecuteForked(Task task, ResolveClusterActionRequest request, A
                 RemoteClusterClient remoteClusterClient = remoteClusterService.getRemoteClusterClient(
                     clusterAlias,
                     searchCoordinationExecutor,
-                    RemoteClusterService.DisconnectedStrategy.FAIL_IF_DISCONNECTED
+                    RemoteClusterService.DisconnectedStrategy.RECONNECT_IF_DISCONNECTED
                 );
                 var remoteRequest = new ResolveClusterActionRequest(originalIndices.indices(), request.indicesOptions());
                 // allow cancellation requests to propagate to remote clusters