After upgrading the vCenter Server to version 9.1, users experience slow UI performance. A banner message is displayed: Could not connect to one or more vCenter Server systems: https://<FQDN_OF_STALE_vCenterServer>:443/sdk
Additional observations:
The following errors are found in the logs:
/var/log/vmware/vsphere-ui/logs/vsphere_client_virgo.log
YYY-MM-SSTHH:MM:SS ERROR vsphere-ui 215 [vc@4413 threadName="wcp-plugins-notifier-1" logger="c.v.v.plugin.notification.WcpServicePluginNotificationsListener" ] Unable to subscribe to vc <FQDN_OF_STALE_vCenterServer> (9a9###c0-f##f-4##9-b##f-6e######35adce), nodeId: dc###c84-1##0-4##d-8##c-d9####70373 java.util.concurrent.CompletionException: com.vmware.vcenter.apigw.api.sso.SsoServiceException: Failed to acquire a token by token...Caused by: com.vmware.vcenter.apigw.api.sso.SsoServiceException: Failed to retrieve an STS client for SSO domain <Stale_Domain>.local(0####09-e##5-4##f-b##0-5d##9c7~CLOUD)
/var/log/vmware/vsphere-ui/logs/apigw.log
YYY-MM-SSTHH:MM:SS ERROR vsphere-ui 8546 [vc@4413 threadName="agw-token-####" logger="com.vmware.vcenter.apigw.sso.impl.SsoServiceImpl" SsoServiceImpl ] Failed to build an STS client for domain <Stale_Domain>.local(0####09-e##5-4##f-b##0-5d##9c7~CLOUD) in 30021 ms com.vmware.vcenter.apigw.api.sso.SsoServiceException: Failed to build a SecurityTokenServiceConfig...Caused by: com.vmware.vcenter.apigw.api.sso.SsoServiceException: Failed to retrieve the STS issuers certificates for domain <Stale_Domain>.local(0####09-e##5-4##f-b##0-5d##9c7~CLOUD) in 30021 ms. (URL: http://localhost:1080/external-vecs/http2/<FQDN_OF_STALE_vCenterServer>/443/sso-adminserver/sdk/<Stale_Domain>)
/var/log/vmware/vpxd/vpxd.log
YYY-MM-SSTHH:MM:SS ERROR vpxd 5528 [vc@4413 sub="cnxmgrLogger" opID="RemoteInstancePreload-75####8c"] Failed to preload instance STS cert; instId: 9#####c0-f##f-4##9-b##f-6e025435adce, host: <FQDN_OF_STALE_vCenterServer> ex: N7Vmacore4Http13HttpExceptionE(HTTP error response: Service Unavailable)
/var/log/vmware/vgfa/ls_configurator/ls_configurator.log shows the stale VMC vCenter is still visible:YYY-MM-SSTHH:MM:SS INFO VGFA 40519 [vc@4413] [CONF_Manager]++++++++++++++ ADDED VC: <FQDN_OF_STALE_vCenterServer>
VCF 9.1
This issue is caused by a stale registration of a decommissioned VMware Cloud (VMC) on AWS vCenter within the on-premises Lookup Service. Although the Enhanced Linked Mode (ELM) relationship was broken or the cloud SDDC deleted, the registration was not properly cleaned up.
On-premises services (VGFA and API Gateway) discover this stale endpoint during startup. When a user logs in, the system attempts to connect to the stale vCenter SDK and SSO endpoints. This triggers multiple 15- to 30-second timeouts, causing cumulative login delays of 75 seconds to 5 minutes.
To resolve this issue, delete the stale registration using one of the following methods.
Take a snapshot of the vCenter Server and ensure a valid file-level backup exists.
World > local > vsphere > configuration > ForeignDomains.ForeignDomains folder containing the stale vCenter registrations.If JXplorer cannot be used, run the following commands from the vCenter Server appliance:
/opt/likewise/bin/ldapsearch -H ldap://localhost:389 -x -D "cn=Administrator,cn=Users,dc=vsphere,dc=local" -W -s one -b "cn=ForeignDomains,cn=Configuration,dc=vsphere,dc=local"
cn value with the ID found in step 1):/opt/likewise/bin/ldapdelete -H ldap://localhost:389 -x -D "cn=Administrator,cn=Users,dc=vsphere,dc=local" -w '<SSO_PASSWORD>' -r "cn=5###d5c-c##a-4##5-b##9-3d#####3ff5,cn=ForeignDomains,cn=Configuration,dc=vsphere,dc=local"