Error when creating Clusters across Availability Zones

I have a ECE setup with 3 availability zones within a network. Within each zone, there are 2 VMs. 1 VM has the proxy and allocator roles while the other VM has the coordinator and director roles.

Whenever a cluster is created which needs to reside on 2 or 3 availability zones, the cluster creation process will fail. The error message is:

Unexpected error during step: [allocate-instances]: [no.found.constructor.steps.allocation.AllocationFailedException: Allocation failed: [CoordinationFailed(org.apache.zookeeper.KeeperMultiException: KeeperErrorCode = NoNode with [CHECK(/, null)), SET_DATA(/clusters/769a5b21eda2420cadf9ee88afc4551a/instances/instance-0000000002, null)), SET_DATA(/clusters/769a5b21eda2420cadf9ee88afc4551a/instances/instance-0000000005, null)), SET_DATA(/clusters/769a5b21eda2420cadf9ee88afc4551a/instances/instance-0000000003, null)), SET_DATA(/clusters/769a5b21eda2420cadf9ee88afc4551a/instances/instance-0000000001, null)), SET_DATA(/clusters/769a5b21eda2420cadf9ee88afc4551a/instances/instance-0000000004, null)), SET_DATA(/clusters/769a5b21eda2420cadf9ee88afc4551a/instances/instance-0000000000, null)), SET_DATA(/services/allocators/ece-region-1b/, null)), SET_DATA(/services/allocators/ece-region-1c/, null)), SET_DATA(/services/allocators/ece-region-1a/, null))])].]

If a cluster is created just on 1 availability zone, then the cluster can be created successfully.
However, if I edit the cluster to run from 1 to 2 availability zones, then the same error message will appear.

How do I address this problem?

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.

Hi @ong - sorry this forum has been so unresponsive - did this ever get resolved?

That's a very strange error ... it would imply that one or two of the 3 zones was not installed correctly (but in a strange way I've not seen before). I'd probably find out which one by using the "maintenance mode" button on each pair of allocators in turn and create a 1 zone cluster, then wipe/reinstall the allocator in question