Site Recovery Manager stopped replicating in one direction
Article ID: 405434
Updated On:
Products
Issue/Introduction
Symptom:
- Reconfigure the current VM replicaiton with NotAuthenticatedFault
- All existing replications to the site are having RPO violations with 'no connection to VR server' errors;
- Failed to set up a new replication.
- Restarting hms service has no effect in resolving the above issue
From vSphere replication appliance (/opt/vmware/hms/logs/hms.log)
2025-07-22 17:28:32.579 INFO com.vmware.hms.gateway.is.impl.LsInventoryServiceFactory [hms-vlsi-client-thread-26959] (..is.impl.LsInventoryServiceFactory$1) [] | (re)login to InventoryService ...OK
2025-07-22 17:28:32.603 WARN com.vmware.vim.query.client.impl.ClientImpl [hms-vlsi-client-thread-26958] (..client.impl.ClientImpl$2) [] | Unable to login to server after retries:2
2025-07-22 17:28:32.603 WARN com.vmware.hms.monitor.host [hms-vlsi-client-thread-26958] (..host.disallowed.DisallowedTagMonitor$PollTask) [] | Unable to compute disallowed object set.
java.util.concurrent.CompletionException: (dataservice.fault.NotAuthenticatedFault)
From vCenter (/var/log/vmware/vpxd-svcs/vpxd-svcs.log)
2025-07-24T09:23:43.491Z [authz-service-6 [] INFO com.vmware.cis.server.authentication.impl.MoSessionManager opId=xxxxxx-xxxx-xxxx-xxxx-xxxxxxx ] User: "VSPHERE.LOCAL\com.vmware.vr-sa-xxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx" already logged in - skipping login request
2025-07-24T09:23:43.496Z
[dataservice-2 [] WARN
com.vmware.cis.authorization.impl.AclPrivilegeValidator
opId=xxxxxx-xxxx-xxxx-xxxx-xxxxxxx IS] User
VSPHERE.LOCAL\com.vmware.vr-sa-xxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx does not
have privileges [System.Read] on object
urn%3Avmomi%3AInventoryServiceTag%3A5ed768a7-bd16-494e-a071-09f0fbf56d49%3AGLOBAL
2025-07-24T09:23:43.516Z
[dataservice-2 [] ERROR com.vmware.cis.core.tagging.vmodl.MoTagManager
opId=xxxxxx-xxxx-xxxx-xxxx-xxxxxxx IS] Failed to query attached objects for given tag : ManagedObjectReference: type = InventoryServiceTag, value = yyyyyyyy-yyyy-yyyy-yyyy-yyyyyyyyyyy, serverGuid = GLOBAL
com.vmware.vapi.std.errors.Unauthenticated:
Unauthenticated (com.vmware.vapi.std.errors.unauthenticated) => {
messages = [LocalizableMessage (com.vmware.vapi.std.localizable_message) =>
{
id
= vapi.security.authentication.required,
defaultMessage = Authentication required
Environment
vCenter 8.0.x and VMware live recovery 9.0.x
Cause
When vSphere replication uses vCenter inventory service to query attached objects for a given tag, it gets NotAuthenticatedFault, as it is using the expired session token and vSphere replication triggers a login after receiving this error, but vpxd-svcs considers it as an already logged in session and skips the login request.This has caused the conflicting message in the hms.log showing above.
Resolution
The issue is known to VMware engineering and planned fix in future release.
The workaround for this issue : (use either one of the below)
a. In vCenter Appliance VAMI page , restart the "VMware vCenter Services" (vmware-svcs) ,
Or from vCenter shell
service-control --restart vmware-svcs
b. Reconfigure the vSphere Replication to its local vCenter.
The goal of the above workarounds is to refresh a a new session token for the communication between vSphere Replication and its local vCenter.
Additional Information
Issue is not reproducible on demand.
In the planned fix, when vSphere Replication receives from vpxd-svcs with NotAuthenticatedFault, it will logout from the previous established session and it will invoke login attempt again to vpxd-svcs once previous logout operation is successful.
Feedback
