Hi guys!
This has been a super frustrating problem, and I'm not sure where else to go with it. I've got an SCVMM2012 R2 Management server set up, and it's managing 4 clusters. 2 of them are all fine, but all of the hosts for the other 2 are in "Needs attention" state, with error 20506. Here's what I've done so far:
Checked the VMM\Admin event log on the hosts, no errors. Checked on the Management server, no errors, just warnings about not being able to copy files to the clusters.
Verified that the Management server computer account is local admin on all hosts.
Verified that the VMM Service account is local admin on all hosts.
Verified that winrm is running on all hosts, and that it's configured.
Verified that the vmmagent service is running all hosts.
Tested winrm against the management server from one of the hosts using winrm id -r:<management server> with no issues.
Tested winrm against the same host from the management server using winrm id -r:<host> with no issues.
Verified that psremoting is enabled on all of the boxes involved.
set the winrm service to its own svchost process using sc config winrm type= own
configured the maxtimeout and maxconcurrentoperationsperuser parameters using the following on one of the hosts:
winrm set winrm/config @{MaxTimeoutms = "1800000"}
winrm set winrm/config/Service @{MaxConcurrentOperationsPerUser="400"}
net stop winrm
net start winrm
net start scvmmagent
Ran MBCA for SC2012SP1 on the Management server, and found a few errors about not having a list of hotfixes for some hosts (working and non-working), one or more volumes having multiple guid paths on hosts (working and non-working), the wmi virtualization store being in an unhealthy state on a few hosts (working and non-working), the VMM SPN being registered incorrectly for the management server (verified the SPN's were registered via setspn -l <management server name> and included in the registry of the management server), and that the winrm service is present, running and healthy on the working hosts of one cluster, but that it's not version 3.0. I don't see anything there that would indicate why the 20506 error is occurring.
Verified via netstat -ano the port that the working hosts are sending over (5985), and verified that the port allows traffic through on both sides by using paping from both management server and host to that port.
Interestingly, the non-working host is not making this connection, which is most likely near the root of my problem. I'm just not sure how to find out why it isn't, and I'm not sure what else to try. I haven't pursued corrupt counters or WMI, because this is affecting all of the hosts on 2 clusters, so the likelihood of it being performance related is small. I've seen plenty of material on this issue, I just haven't been able to find anything out there that has worked for me so far. Does anyone else have any other ideas?