VMware Cloud Community
muxonarization
Contributor
Contributor
Jump to solution

Appliance 6.0U3 to 6.5a fails with Internal Error during Export of EAM

Hello,

I get an update error when upgrading vCenter Virtual Appliance 6.0U3 to the latest 6.5A.

The installer tells me "Internal error occurs during Export of vSphere ESX Agent Manager."

In the downloaded Export_com.vmware.eam_$DATE.log i find the following:

INFO service_manager Service vmware-eam reported status started. Expected status started

INFO service_manager Service vmware-eam: STARTED

INFO eam Successful start of EAM service.

INFO eam Going to access: http://localhost:80/eam/charset.jsp

WARNING eam vmware-eam service is not fully started, waiting till get up and running. Sleeping for5seconds. Error: HTTP Error 503: Service Unavailable (Failed to connect to endpoint: [N7Vmacore4Http16LocalServiceSpecE:0x7fd7aa519970] _serverNamespace = /eam _isRedirect = false _port = 15005)

WARNING eam vmware-eam service is not fully started, waiting till get up and running. Sleeping for5seconds. Error: HTTP Error 500: Internal Server Error

the last line repeats several dozen times, followed by

ERROR __main__ Upgrade Phase 'eam:Export' failed. Exception: expected string or buffer

at the near end of the logfile.

I already tried the steps from KB2147898 - which is for Windows vCenter btw. - to no avail.

There are no log entries regarding heap issues or a custom Port.

I already restarted the original VCVA several times, restarted vmware-eam via the VCVA shell and immediately tried to migrate afterwars without any success.

Steps taken before the upgrade:

- migrated vShield 5.5.4.2 to NSX 6.2.5 and then to 6.3

- uninstalled VUM so I don't have to migrate it, plan was to start using the integrated VUM after the migration.

- removed vcIntegrity Extension

- timesyncd hosts to internal NTP, configured VCVA to sync with host, times are fine on all hosts and the appliance

Desperate steps taken after the error occured:

- set a new and easy password for SSO and appliance-root

- using only IP adresses instead of hostnames in the migration assistant

- tried the upgrade on my 6.0U2 homelab, no vShield/NSX there, uninstalled VUM, same error

- tried editing "bypassVumEnabled" per KB2053782, but the value keeps saying "unset" after committing the change, i suppose that's because there's no more vum

Is it because of the VUM uninstall/extension removal?

What else can i do?

Setting up a fresh VCVA65A is out of question, since we use VDP 6.1.3 and a new vCenter would most definately not allow for VDP appending to existing backups, thus creating new restore points and great demands in storage capacity.

Any hints are greatly appreciated!

regards

Max

Reply
0 Kudos
1 Solution

Accepted Solutions
ashwinick
VMware Employee
VMware Employee
Jump to solution

as per vcenter upgrade matrix VMware Product Interoperability Matrixes from 60U3 to 65a we do not have any support .

View solution in original post

Reply
0 Kudos
4 Replies
muxonarization
Contributor
Contributor
Jump to solution

Update:

Deployed a second VCVA 6.0U3 on my hosts, logged in once to make sure everything's green and immediately started the upgrade to 6.5a.

Mind this is a completely fresh and naked appliance with a static IP, no hosts or solutions integrated.

Guess what, EAM error, logs are all the same.

As the appliance is a VM, i suppose the underlying host hardware does not play any role in this issue, right?

My network:

production VCVA6U3 is 192.168.0.5

hosts are 192.168.4.11-13

installer is 192.168.2.160

network is /21, so it's all in the same subnet, no firewalls between hosts/vcva/installer.

What could be the culprit?

Reply
0 Kudos
ashwinick
VMware Employee
VMware Employee
Jump to solution

as per vcenter upgrade matrix VMware Product Interoperability Matrixes from 60U3 to 65a we do not have any support .

Reply
0 Kudos
muxonarization
Contributor
Contributor
Jump to solution

correct, i just upgraded a fresh 6U2 appliance to 65A without issues on the very same environment.

I could've sworn i tried that with my homelab already, also getting the EAM error.

Well, then i will just wait for support / migration path.

The way i read the matrix before the upgrade was that U3 was so new as there's no column for it, you VMWare guys just forgot to update the matrix 🙂

And there was some manual stuff recommended to do before the upgrade, which was corrected in U3. So i thought - go U3 first, then 65A, less work.

Thanks for your answer.

Reply
0 Kudos
ryansnyder
Contributor
Contributor
Jump to solution

We have a lab environment and are seeing the same issue. We were trying to go from 6.0u3 VCSA with external PSC to 6.5b which just came out. We are seeing the same problems. Apparently don't update to 6.0u3 if you plan to move to 6.5 soon.

Reply
0 Kudos