Pages

Wednesday, March 4, 2015

Skype for Business Readiness Series (4/15)

Today I joined the fourth session of the Skype for Business Readiness WebCast Series and the topic for today was the Manageability Improvements Overview.

The Webcast was recorded and will be made available offline within a month from today and also the slides will be available for download one month post the session.

Lynn Roe covered the in-place upgrade procedure and discussed the traditional migration which apparently is referred to as "swing migration." A full in-place upgrade was demoed in power point and in parts in a real environment.


The multi-homing issue, when a Front-end server have more than one IP, is not  solved in S4B. Lync 2013 have some problems and Microsoft recommends using a single IP address on any Front-end Server. Microsoft are considering releasing a KB article to list the workarounds so that dual NICs/IPs can be used.

A graphical user interface for CLS were demoed and it looked a lot like the good old OCSlogger.exe, and a demo with a SQL AlwaysOn fail over finished today’s session.

Here comes the QnA, 32 new answers:


New features

What will be the best practice / recommended approach: In-Place Upgrade or  Swing Migration?
Depends on your business requirements and hardware availability.

Is it possible to still use the "swing migration" method to introduce S4B, e.g. setup a new S4B pool, and then migrate users to the new pool?
Yes, we can still do swing migration method in S4B.

Can we do a swing migration from Lync 2010?
Yes, swing migration is the option to migrate from Lync 2010 to a S4B pool.

What is the upgrade order we have to follow when doing the upgrade?
Typically: 1) Front End Pools / Standard Edition Servers, 2) Directors, 3) Mediation, Persistent Chat, 4) Edge

Assuming multiple Lync 2013 pools and we are doing an in-place upgrade, should we upgrade the pool hosting the CMS first?
No, the pool hosting CMS does not need to be upgraded first.

Is there any fallback scenario for in-place upgrade, if something goes wrong?
As long as you maintain a copy of the topology before it was upgraded and a backup of the backend databases, you can uninstall/remove S4B and reinstall Lync Server 2013.

For an in-place upgrade, how long is the outage?
This epends on several variables like the number of pools, number of FEs in a pool, DB size, SE vs EE.

Are all meeting reservations and meeting content preserved by the in-place upgrade process?
Yes.

On a standard edition server, at which stage of the in-place upgrade are the users disconnected?
As soon as the in-place upgrade starts on the standard edition server.

Do we have to unpair paired pools first before an in-place upgrade?
No, no need to unpair the pools.

We have an active pool and a paired backup pool that contains no users unless we are failed over to it. Would it be reasonable / supported to update the backup pool first?
Yes, we can update the pools in any order we prefer.

What Cumulative Update is required on Lync 2013 to start upgrading?
Minimum of the August 2014 cumulative update for Lync Server 2013, but  strongly recommend is the latest public update.

Are there pre-checks to ensure all required prerequisites are in place before launching an upgrade?
Yes, the in-place upgrade process has a prerequisites check that will block install until all prerequisites are met.

Will Skype For Business 2015 work with OCS 2007 R2 edge servers?
No, OCS 2007 R2 is not supported with Skype for Business Server.

If we have a pool with three servers in a pool, does all three servers have to be upgraded at the same time for the pool to come online again?
Yes, all servers in a pool have to be upgraded at the same time.

We have Lync 2013 in Hybrid mode can we stay at Lync 2013 on premise when the Office 365 version becomes Skype for business online?
The hybrid environment for already moved users should be supported but if we move an on-prem user to online (which runs S4B) then there might be loss of user data because of different server/service versions. There is a warning shown in the control panel when you perform this step.

Is Start-csPool like reset-fabric?
Yes, Start-CsPool does a similar job. The cmdlet is also better adept at handling failures and startup services.

What does Windows fabric v3 provide that v2 does not?
Windows Fabric v3 handles rebuilds on slower networks much better, better resiliency of services when some fabric services were offline, fixes for crashes that occurred during replica movements, ability to enable remote copy of winfab logs, ability to control the size of winfab traces, better handling of internal communication.

Do trusted applications e.g. OWA/IM, Watcher Node add-ons need to be uninstalled prior to publishing the topology for S4B?
No, no changes needed on Trusted applications.

Have CLS (Centralized Logging Server) been improved in S4B?
There are a number of CLS fixes, including a new CLS UI tool called CLSLogger.exe.

Can we do in-place upgrade for a Lync 2013 pool that uses a SQL 2008 R2 Back-end?
Yes.

If we do an in-place upgrade of a pool using a SQL 2008 R2 Back-end, can we upgrade the SQL Back-end to 2012 later to take advantage of the SQL always on feature?
Yes.

Is SQL Always On required for in-place upgrade to work?
No, SQL always On is not a requirement for in-place upgrade.

Can we change into SQL AlwaysOn configuration during the in-place upgrade process?
No, but it is supported changing the SQL High Availability model AFTER the pool has been upgraded to S4B.

Are the SQL Enterprise Edition or SQL-Enterprise CALs needed for a 2-node Always-On-System, or is Standard sufficient?
Depends on if Availability Groups (Enterprise Edition) or a two-node Failover Clustering Instance (Standard Edition) is being used.

If we want to migrate from SQL mirror to SQL AlwaysOn. Do we just make SQL mirror instance offline and copy the database to SQL Avalability group instance?
Migrating from SQL Mirroring to SQL AlwaysOn requires that you break the mirroring relationship first.

Can we do the SQL Always-on failover with PowerShell from a S4B server?
There currently is no database failover cmdlet for SQL AlwaysOn Availability Groups in the S4B Management Shell. However, there may be a cmdlet to do this in the SQL PS module.

In order to complete the in place upgrade, do the local servers need to run SQL 2012 SP1?
Skype for business supports a SQL Server 2008 R2 SP2 backend, but SQL 2012 Express SP1 is needed on the local instances. The In-place upgrade will update the databases and Microsoft recommends using Test-CsDatabase to check that all databases are updated.

How long time does the initial publish after an upgrade take in larger environments?
It depends on how large your file stores are (content-wise). Following the initial publish of the upgraded topology, the Enable-CsTopology process will re-ACL all of the contents of the associated file store. Depending on the number of items, it can take anywhere from a few minutes to a few hours.


Unchanged features

Is database mirroring still supported in SQL 2014?
Yes it is still supported in SQL 2014, TechNet states that it will be removed in a future version.

Can a S4B pool be paired with a Lync 2013 pool during a coexistence phase?
No, pair the pool with a similar version.

Having a pool on Windows 2008 R2, we want to leverage Windows fabric 3.0, can we take one Front-end out of the pool, upgrade it to Windows 2012 and add it back to the same 2013 pool again, ending up with a mixed OS pool?
No, mixed OS pool is not supported.

No comments:

Post a Comment