|Article Number: 540743||Article Version: 2||Article Type: Break Fix|
ViPR Controller,ViPR Controller Controller 3.6 SP2
The user is unable to add a new host to a stretched cluster (isolated fabrics) in a VPLEX distributed environment. VIPR Controller logs
vipr1 vipr1 controllersvc 2020-01-21 12:59:12,114 [RMI TCP Connection(15608)-10.10.10.10] ERROR ComputeSystemControllerImpl.java (line 750) addHostToExport caught an exception.
com.emc.storageos.computesystemcontroller.exceptions.ComputeSystemControllerException: Initiator Initiator(Protocol:FC, Node:20:00:00:00:00:00:00:00, Port:20:00:00:00:00:00:00:00, Host Name: serverName.domain, Cluster Name: ),Initiator(Protocol:FC, Node:20:00:00:00:00:00:00:00, Port:20:00:00:00:00:00:00:01, Host Name: serverName.domain, Cluster Name: ) does not have connectivity to volumes in Export Group ExportGroupName (urn:storageos:ExportGroup:fe1fb5b0-8e37-4a8c-a495-6ae252c5aaa8:vdc1)
VIPR Controller logs
The VIPR Controller logs show that when verifying connectivity to the volumes in one of the cluster export groups, VIPR-C does not find the host initiators and ExportGroup storage ports in the same fabric, which causes the Update Export Group task to fail.
In this environment, there were 2 export groups for the cluster (both VPLEX distributed).
The ExportGroup1 was configured with a VirtualArray of site 1 & AltVirtualArray of Site 2.
The ExportGroup2 was configured with a VirtualArray of site 2 & AltVirtualArray of Site1.
The host being added to the cluster is located at Site 1.
Volumes in ExportGroup1 are both VPLEX distributed & VPLEX local and there is connectivity between these volumes (Site1) and the new host (Site1) as all volumes are in the same Fabric as the new host.
Volumes in ExportGroup2 are only VPLEX local and therefore there is no connectivity between these volumes (Site2) and the new host (Site1). As a result VIPR-C fails the connectivity verification and as a result, the ExportGroup update fails.
1.Verify that the ExportGroup causing the error is a valid ExportGroup with valid volumes.
If the volumes in the ExportGroup are stale, then the ExportGroup can be deleted and re-run the add host ActionableEvent/task.
2. If the ExportGroup is valid, but only contains VPLEX local volumes (in a VPLEX distributed environment) then provision a new placeholder VPLEX distributed volume to the erroring ExportGroup first.
VIPR Controller needs connectivity to exist between the new host and at least one volume in the ExportGroup.
The new VPLEX distributed volume will have copies on both sides of the VPLEX cluster (Site 1 & Site 2) therefore VIPR-C can validate connectivity from the new host (Site1) to the copy of the new volume on Site1.