Replica View Connection server install rollback

[Zomig]

Replica View Connection server install failed with error code 1603.

File vminst.log

adamInstUtil: 01/22/20 12:16:35 Begin Logging

adamInstUtil: 01/22/20 12:16:35 --- CA exec: VMAdamReplicaImportCMS

adamInstUtil: 01/22/20 12:16:39 ERROR: Cannot bootstrap CMS: Cannot connect to any node, attempts=1, last-error=Channel timeout/abort during auth

adamInstUtil: 01/22/20 12:16:39 End Logging

File vmmsi.log

Acción 12:16:35: VM_AdamReplicaImportCMS.

MSI (s) (78:40) [12:16:35:919]: Executing op: CustomActionSchedule(Action=VM_AdamReplicaImportCMS,ActionType=1025,Source=BinaryData,Target=VMAdamReplicaImportCMS,)

MSI (s) (78:FC) [12:16:35:934]: Invoking remote custom action. DLL: C:\Windows\Installer\MSIEC91.tmp, Entrypoint: VMAdamReplicaImportCMS

CustomAction VM_AdamReplicaImportCMS returned actual error code 1603 (note this may not be 100% accurate if translation happened inside sandbox)

La acción finalizó 12:16:39: InstallFinalize. Valor de retorno 3.

Could you please help me?

There is no info on Internet about this. When the installer finish then rollback.

[Alex_Romeo]

HI,

read this KB:

VMware Knowledge Base

ARomeo

[Zomig]

Hi thansk for your response AlessandroRomeo68

Tthat kb of vmware does not apply to my case. However, I have run the command on the only connection server I have and on the replica server that does not let me install. With no results. The installation of the replica server keeps rollingback.

Both VM dont have Windows firewall and (on my opinion) the replica server get successfully the ADAM database from the connection server.

The problem are on this logs. For one reason it keeps falling.

[Jon_Holloway]

Can you double check that port 32111 between the replicate and standard server is not blocked? The error you are seeing can occur if that port is blocked. The replication would be successful over port 389, but exchange of the cluster master secret (CMS) would fail if port 32111 is blocked.

Thanks

Jon

[sjesse]

Check all the ports, Jon mentioned 32111 but this post says 135 caused this issue. Its most likely a physical firewall, the connection server software should configure the windows firewall

Replica Connection server install failed with error code 1603 - 7.9.0-13956742

In the other post there is a like to the required ports that are needed  for a sucessful enviornment.

[Zomig]

Hi,

I make the proofs with no firewall on the connection server primary and secundary. There are both virtual machines on the same esxi.

With our firewall rules in both connection server (u can clic on the photo for to see all the ports):

The instalation wizard freeze here:

"Configuring the instance of replica vmwareVDMDS directory services"

As i said without firewall in both virtual machines the instalation continues but in the last step it roll back with the error.

Does anyone know what is happening?

sjesse​ AlessandroRomeo68​

[Jon_Holloway]

Hi Zomig, have you tried installing a replica Connection Server on a fresh Windows Server, that you haven't previously had Connection Server installed on or attempted installation on? 

-Jon

[sjesse]

Your showing the windows firewall, is there a firewall in the network? Just because they are in the sam esxi, if there not in the same subnet the traffic leaves the esxi server and comes back.

[TJJ25]

Hey Zomig,

I encountered the same issue today while doing free trials of Horizon 7.
Did you manage to find a solution?

In my case it's definitely not a firewall problem.

[Zomig]

No, i dont have the solution yet...

[Shreyskar]

Hi Zomig

Error 1603 is a generic error code and doesn't tell us anything specific why the installation failed.

From what I can interpret from the log snippet is , it is failing at the time of repilcating AD LDS with primary connection server (CustomAction VM_AdamReplicaImportCMS ). This may happen to mismatched OS version due to which AD LDS (ADAM) schema corruption happens and it won't replicate.

If you are looking into horizon install logs, search with keyword 'return value 3' and then track down each event above that.

Action plan:

> Make sure primary and replica CS are running same OS otherwise ADAM database replication will fail due to schema mismatch

> Take a backup of ADAM database on existing connection servers (See VMware Knowledge Base ).

> Take a snapshot of existing connection servers.

> Uninstall replica connection server from programs and features if it is there.

> Uninstall 'AD LDS' component from replica connection server.

> Remove the reference of replica server from the remaining "good" connection server via the following command.

vdmadmin -S -r -s <replica-server-FQDN>

Note: You need to run above command on a good working connection server.

> Delete 'VMware VDM' reg hive from below location on replica server:

HKLM\Software\VMware\VMware VDM

If delete operation fails, right click reg hive > permissions > advanced > check the box 'replace all child object permission...' > Try to delete it now.

> Rename or delete the folder C:\Program files\VMware\VMware View

> Open certificate manager (local machine) and delete all certs from 'VMware horizon view certificates' location. These will be recreated during installation. Do not delete any cert from 'Personal' store.

> Reboot replica server.

> Install replica server again pointing to primary CS.

[jhanekom]

I vrealise this is an old thread, but just thought I'd pop in and say that I had what appeared to be same problem (at the very least, the same error messages, which led me to this thread.)

I had tried several times to uninstall/re-install the replica connection broker with various procedures as well as doing a re-install with LDAP restore on the last remaining connection broker, but the one that worked in the end was this one.  On a hunch, it would either be deletion of the reg key or the certificates that helped - or both.

Thanks @Shreyskar!

[future2000]

I also note this is an old thread. However I was having the exact same issues with Horizon View 8.4. It seems the removal of the replica from the initial connection servers Db does not happen with an uninstall. Removing this replica instance from the primary database with vdmadmin -S -r -s <replica-server-FQDN> command solves the problem and allows the installation of the replica to complete if it failed previously.

[Kelvin_Koh]

this solution helped me. thank alot

[Helliouse]

Sorry to resurrect this post. It has been the only information that has helped to trouble shoot this issue.

I am installing Version 8.7.0-20649599 and ran into this same issue, read https://kb.vmware.com/s/article/76941 and followed it. Performed a full uninstall on all my servers. No connection issues between servers (netstat -ano and Test-NetConnect [fqdn] -Port [####]) to verify. Tried all the suggested solution in the posts above to no avail. 

What did work?

Exporting the vdm.ec certificate from the master Connection server, and importing it into the replica server.

If any one knows what process if failing to transfer the certificate between the servers to help me troubleshoot the failure and come up with a better resolution that would be great.

Hope this helps someone else.