Of late, I have been working on installation of SFB (Skype for business server). I was setting up persistent chat recently following the documentation and went through multiple issues.

Using topology builder – created a persistent chat pool & selected option to create with single server. Named the pool as pool02.mydomain.com and server name was server-skype2.mydomain.com. Published the topology,

Issue 1 : the deployment on the server failed with error that the server was not part of the topology.

Essentially, pool & server name are different required a topology with persistent chat pool with multiple servers. I went ahead & deleted the persistent pool & recreated with option to create with multiple server. With the change, deployment on the server went through Install Local configuration store.

Issue 2 –  The tick mark against the Install Local configuration store step in the deployment wizard was missing even after the step completed without any failure.

I had missed creating pool02 DNS AAA record. I also learnt that apply latest CU for the release is recommended before diagnosing any installation issue. Error at this stage essentially meant central configuration store is not replicated to the server or published topology did not have appropriate entry for the server.

Downloaded the latest CU as per the hit.  Used this handy powershell command to stop all the services as required by the update installer.

// Stop all the services
> get-cswindowsservice | stop-cswindowsservice

// After applying the update - start all the services
> get-cswindowsservice | start-cswindowsservice

// to query the central management store replication status on all server nodes
> Get-CsManagementStoreReplicationStatus
UpToDate           : True
ReplicaFqdn        : myserver.mydomain.com
LastStatusReport   : 12/21/2017 5:31:12 PM
LastUpdateCreation : 12/21/2017 5:31:09 PM
ProductVersion     : 6.0.9319.0

UpToDate           : True
ReplicaFqdn        : myserver02.mydomain.com
LastStatusReport   : 12/21/2017 5:31:12 PM
LastUpdateCreation : 12/21/2017 5:31:09 PM
ProductVersion     : 6.0.9319.0

// UpToDate : False means configuration is not getting replicated. 
// Application of CU & restarting of the services on the server hosting the master (replication) & this server fixed the replication issue
// UpToDate : True means configuration replication is ok.

Issue 3 : mgc database is not created in the sql server instance in the backend.

The topology had next version of the SFB server essentially N, N-1 versions. In such topology, deployment of persistent chat pool under N-1 node does not create mgc database when topology is published by the topology builder. It has to be created using the PS command install-csdatabase on the N-1 version server using (N-1) version skype server powershell

> Install-CsDatabase -ConfiguredDatabases -SqlServerFqdn sql.mydomain.com -ForInstance mysqlinstance 
// this created the mgc database in the backend sql instance

Prior to doing this – I have also followed the step (1) on this page to ensure that networking ports are open (firewall) & computer network discovery is enabled. I had just disabled firewall to deal with firewall settings later. One essential step is to ensure the logged on user on the skype server is sqlserver instance sysadmin.

Issue 4 : persistent chat service failed to start on the server.

// Error in the lync server application event log
Skype for Business Server 2015, Persistent Chat could not start due to the following exception:
 at 
Microsoft.Rtc.Internal.Chat.Server.ServerCommon.Exceptions.StopServerException: CmdID: 871d1cfd-bb20-4c25-a1cb-c4e5ee7fb956 The server could not restore db connection within the allowed time (00:10:00) using connection string: Data Source=sql2.mydomain.com\mysqlinstance;Initial Catalog=mgc;Integrated Security=SSPI. ---> System.Data.SqlClient.SqlException: Login failed for user 'mydomain\myskypeserver$'.    

// myskypeserver$ is the computer account & it did not have permission to the mgc database

I found using SS management studio that mgc database had connect grant for the group `mydomain\RTCComponentUniversalServices`. Skype setup adds the computer account to this AD group during setup. For the server – it was missed. I added the mydomain\myskypeserver (computer account) to this AD group manually. With this, the persistent chat service could start and did not fail with the above error after 10 minutes.

Issue 5 :  skype client continues to show warning Your chat room access may be limited due to an outage

<I am working on this>