Symptoms
The vCenter Server does not automatically deploy vCLs after attempting retreat mode due to an agency in yellow status.
wcp.log remain in the deletion and destroying agent loop. There are no entries to create an agency.
2023-06-24T15:30:13.28Z debug wcp [dataprovider/cluster_agencies.go:96] [opID=vCLS] Successfully deleted from the database record (domain-c24137)
2023-06-24T15:30:13.28Z info wcp [eamagency/retreat.go:58] [opID=vCLS] Successfully destroyed agency, retreat operation completed
There is no relevant information in eam.log
In the eam_api-x.log you may see an error similar to the following:
2023-06-20T14:29:13.130Z | INFO | vlsi | LocalizationFilter.java | 108 | API COMPLETE: Facade for ClusterAgent(ID: 'Agent:3828c184-2c7a-4e15-9e2f-6059e3eb17a1:null').queryRuntime[opId=668781274, sessionId=2A280743]. Result:
eam.Agent.RuntimeInfo {
issue = <unset>,
goalState = 'enabled',
entity = 'Agent:3328c105-2c3a-4e15-9e2f-6059e3eb17a1:null',
status = 'yellow',
vmPowerState = poweredOff,
receivingHeartBeat = false,
vmIp = <unset>,
esxAgentResourcePool = <unset>,
esxAgentFolder = <unset>,
installedBulletin = <unset>,
vmHook = <unset>,
agency = 'Agency:10458969-7b9e-4ac9-b14e-45f6e2uyt5nvvc:null',
vm = <unset>,
installedVibs = <unset>,
vmName = '',
host = <unset>,
Purpose
Assist customers to solve this issue.
Cause
There are EAM agencies in yellow status.
Impact / Risks
High because the DRS does not properly work causing an imbalance of cluster resources.
Resolution
Check if there are any agencies in yellow color. Restarting EAM usually fixes the problem.
service-control --stop vmware-eam && service-control --start vmware-eam
After restarting the services follow the steps 1 through 8 from https://docs.vmware.com/en/VMware-vSphere/7.0/com.vmware.vsphere.resmgmt.doc/GUID-F98C3C93-875D-4570-852B-37A38878CE0F.html or https://kb.vmware.com/s/article/91890?lang=en_US to enable "Retreat Mode".Wait for 1 to 2 minutes to make sure the VCLs VM's were removed from the cluster.Disable retreat mode by changing the config.vcls.clusters.domain-c<number>.enabled value to "true".
Related Information
https://kb.vmware.com/s/article/91890?lang=en_US
