For the dinner, the UK team chose a fabulous Italian restaurant called Bacco (www.bacco.de/english/restaurant/restaurant.html) which I’d definately go back to and hosted a great evening… many thanks to Stuart, Gareth, Neil, Alex, etc. from Microsoft UK. 

We were also joined by Allen Stewart & Rajesh Dave from corp.  Allen is Principal PM for Windows Server and Raj is a PM for Windows Hyper-V.  Both very interesting & incredibly knowledgable guys with deep understanding across a wide range of topics (and not just Microsoft!).
I pestered them for info on Hyper-V thin provisioning of memory and whilst they couldn’t confirm anything as we all said ‘we live in hope!’ ;-) 

…as for the  night, I’d been invited to the 1E TechEd Europe party at Spindler & Klatt www.spindlerklatt.de - an uuber trendy restaurant/club in East Berlin frequented by the likes of Angelina, Clooney, and now Cook! 

What a great party and many many thanks to the team at 1E (www.1e.com).  Did I mention I was the 4th member of the business in the founding year?  (yes I probably did & several times.. lots to drunk! ) We went our separate ways in 1999, oh for a slice of that now… anyway, moving on! 

Seriously though hats off to Samir, Mark, and Phil – they have built a company that knows how to throw a great party (regarded as the best at TechEd), and a team of very bright, talented people who have a lot of respect for the company and its founders. 

Ouch my head is pounding!  time to go to sessions, starting with… 

ITS211 Keeping Your CIO Happy: Microsoft Office SharePoint Server 2007 SLA Scorecarding with Operations Manager 2007 and SQL Server 2008

Presenters: Gordon McKenna, Sean Roberts, www.inframon.com

Thu 11/12 | 10:45-12:00 | London 2 – Hall 7-1b

Learn how you can create CIO level SLA scorecards in SharePoint Server 2007 for Microsoft System Center Operations Manager 2007 using some of the new features in Microsoft SQL Server 2008 Reporting Services and to create Executive SLA views of your Operational Environment. The session looks at why these types of views are important to many companies, what impact this can have on your business, and what simple steps you can take to achieve very effective, high-level executive views of everything from performance and availability of your key LOB services and applications, whether important SLAs and KPIs are being achieved and whether your IT department is meeting the day-to-day needs of your business. The key demos in this session take you through the steps you need to implement effective business scorecarding in SharePoint Server 2007 using key metrics collected in the Operations Manager 2007 Datawarehouse based on “real-world” experiences gained from the field. After attending this presentation you will have a good insight into how CIO Scorecards can help you add value to your Operations Manager deployments, helping you to show real value to your executives.

Tip – to remove parameter data from Ops Mgr reports imported into a SharePoint webpart, suffix the url with &rc:Parameters=collapsed

Cracking session from Gordon & Sean on how to try and keep your CIO happy (if that’s possible! )

blog Daniel Savage

Service Level dashboard – free solution accelerator dashboard on Microsoft 

SVR401 & 402 DirectAccess Technical Drilldown, Part 1 of 2: IPv6 and Transition Technologies + Part 2 of 2: Putting It All Together

 

Presenter: John Craddock (www.xtseminars.co.uk)

Thu 11/12 | 13:30-14:45 | Helsinki – Hall 7-2a

Take a sprinkling of Windows 7, add Windows Server 2008 R2, IPv6 and IPsec and you have a solution that will allow direct access to your corporate network without the need for VPNs. Come to these demo-rich sessions and learn how to integrate DirectAccess into your environment. In Part 1 learn about IPv6 addressing, host configuration and transitioning technologies including 6to4, ISATAP, Teredo and IPHTTPS. Through a series of demos learn how to build an IPv6 Network and interoperate with IPv4 networks and hosts. In Part 2 we add the details of IPSec, and components that are only available with Windows 7 and Windows Server 2008 R2 to build the DirectAccess infrastructure. Learn how to control access to corporate resources and manage Internet connected PCs through group policy. Part 1 is highly recommended as a prerequisite for Part 2.

John Craddock is an extremely talented AD/identity expert, and deeply technical across many other fields – in this case IPv6 & DA.

I was also lucky enough to have a drink with John and my old Microsoft PSS chum Paul Duffy on Monday night at the cleverly named hotel ‘Berlin Berlin’.

John is a genuine international industry expert and a thoroughly nice bloke with it!   Paul, another ‘genie-I’ went on to become PM for Office Communicator and knows a thing or ten about OCS amongst other subjects to a deep level.  This probably explains why these two know each other!

Anyway, back to the session plus my own notes, links, etc.

Gems & Tips

- be careful, not all apps will be compatible – test!
- to be native will likely mean new network gear, is new network layer (layer 2 unchanged)
- hex is back!  use of double colon notation, but can only be used once per address
- cannot mix with ipV4 mask bit notation
- host derived with mac address which has privacy issues, Win7 & R2 generate random based on interface, can be disabled (revert to mac based) with netsh interface ipv6 set global randomizeidentifiers=disabled

- route print -6 will show IPv6 route table

- ::1 is IPv6 loopback

- if you have a registered IPv4 address then you automatically have an IPv6 address on the 6to4 network

6to4 http://en.wikipedia.org/wiki/6to4 states 6to4 performs three functions:

1. Assigns a block of IPv6 address space to any host or network that has a global IPv4 address.
2. Encapsulates IPv6 packets inside IPv4 packets for transmission over an IPv4 network using 6in4.
3. Routes traffic between 6to4 and “native” IPv6 networks.

- you need to manually unblock ISATAP entry in DNS which can be done via the registry or command line, e.g. 

C:\>dnscmd /config /globalqueryblocklist wpad 

Registry property globalqueryblocklist successfully reset.
Command completed successfully. 

ISATAP is a huge subject in it’s own right, the Intra-site Automatic Tunnel Addressing Protocol Deployment Guide is available at http://www.microsoft.com/downloads/details.aspx?familyid=0f3a8868-e337-43d1-b271-b8c8702344cd&displaylang=en 

Putting it all together..

- Check tunnel endpoint authentication using ‘klist’ to list Kerberos data
- Use NRTP to direct DNS queries to a specific server for a particular names space (view using ‘netsh namespace show effectivepolicy’)
- PKI needs to be right as certificates are the foundations
- you must publish the revocation list
- NLS (Nework Location Server) is just a https website accessible from the DA server, e.g. nls.corp.example.com
- if it doesn’t work, it could be a couple of days troubleshooting! 

If you’re thinking of setting this up in a virtual lab, I also took note from Allen Stewart’s blog at http://blogs.technet.com/wincat/

…if you’re planning to virtualize your lab environment on Hyper-V, you should ensure you’re using Legacy Network Adapters for the child partition where you’re running the DAS. Using the default synthetic NICs is OK for all the other resources in the test lab, but for the DAS itself, it’s important to have both the Internet and Corpnet NICs as legacy ones, to ensure proper passing of traffic between both sides of the DAS. If you use the default synthetic adapters, you may end up in a situation where traffic doesn’t properly flow from the outside to the inside, even though all your IPsec, 6to4, Teredo, and IP-HTTPS settings are correct. Basically, you’ll be in a situation where connectivity will fail at a basic level, with you not even being to successfully ping the internal DNS server using its ISATAP address.If you’ve already built your lab on Hyper-V using the synthetic adapters, the fix is pretty simple. Just replace them with legacy ones, reconfigure the IP addressing as specified in the guide and rerun the DirectAccess wizard, again supplying all the information specified in the guide. After doing so, all your traffic should flow properly.

- Thanks Allen!

DAT312 All You Needed to Know about Microsoft SQL Server 2008 Failover Clustering

Presenter: Gopal Ashok

Thu 11/12 | 17:00-18:15 | London 3 – Hall 7-1b

There are major architectural changes in SQL Server 2008 for failover cluster setup and management, geared towards increased reliability and high-availability. To learn all the benefits and changes, attend this session for a comprehensive overview direct from the product development group. We cover SQL Server 2008 failover clustering setup, underlying Windows Server cluster and how SQL Server uses it, what’s new in SQL Server 2008 for failover clustering, differences from previous versions of SQL Server and future directions. This includes details of SQL Server 2008 failover clustering setup operations together with demos to illustrate the new setup.

- new features
- applications need retry mechanisms built in to provide seamless failover
- no longer have to take down the cluster to upgrade, supports rolling upgrades 

Want to deploy stretched clusters?  lots do.  As in separate geo-redundant clusters, not separate nodes e.g. 

Stretched SQL Clusters or the doodles of an artist?

- sql 2008 failover clustering install breaks on windows server 2008 R2 and needs to be slipstreamed with SP1 (If only we knew this last weekend!)
(slipstreaming is incorporating patches into the installation media to effect a higher level of install base over RTM – Microsoft tend to do this but not always quickly!)
see http://blogs.msdn.com/psssql/archive/2009/03/17/how-to-fix-your-sql-server-2008-setup-before-you-run-setup-part-ii.aspx for more info
- during upgrades to a 2-node cluster there will be a period of time when you are exposed to node failure, and must not have a failover attempt for fear of corruption.  removing the node from the cluster owners will stop premature attempted failover. 

Further Microsoft resources.. (will add others also) 

      SQL Server ^® 2008 Failover Clustering White Paper: http://sqlcat.com/whitepapers/archive/2009/07/08/sql-server-2008-failover-clustering.aspx 

      Recommended  Books Online  Doc Refresh #7 (May, 2009), or later: http://msdn.microsoft.com/en-us/library/ms130214.aspx 

      Failover Clusters – Getting Started: http://msdn.microsoft.com/en-us/library/ms189134.aspx 

      Rolling upgrade process and best practice: http://msdn.microsoft.com/en-us/library/ms191295.aspx 

      Maintaining a Failover Cluster: http://msdn.microsoft.com/en-us/library/ms178061.aspx 

      Setup command line usage: http://msdn.microsoft.com/en-us/library/ms144259.aspx 

      Configuration.ini file usage: http://msdn.microsoft.com/en-us/library/dd239405.aspx 

Not the best day for me in terms of TechEd objectives (i.e. attending learning sessions, etc.) with the first post of the day saying ‘decisions decisions… for sessions just 09:00-10:15′ as struggled to choose between:

DAT302 Top 10 Best Practices for Microsoft SQL Server 2008 Analysis Services
or
MGT11-IS Get Virtualized with Microsoft System Center Essentials!
or
OFS322 Overview of Social Computing in SharePoint 2010
or
SVR207 Windows Server 2008 R2 File Classification Infrastructure: Managing your file data more effectively.
or
SVR319 Multi-Site Clustering with Windows Server 2008 R2

Ended up doing none of the above, but did have a productive breakfast meeting with Stuart Leddy, UK Windows Server Product Marketing Manager.  Stuart has been heading up the Windows Server 2008 R2 UK EAP activities that we have been involved in with Ascom Network Testing (http://blog.thefullcircle.com/2009/11/05/the-full-circle-secures-ascom-network-testing-for-windows-server-2008-r2-early-adopter-program/)

The day turned into going from one full session to the next, and walking back and forth for what seemed like miles in between!
my TechEd tip for the day - for popular sessions get there 10 minutes before they start!

Eventually did get into a 1st choice session I wanted at 12:20…

MGT03-DEMO Introduction to Microsoft System Center Essentials 2010

SVR307 Security Best Practices for Hyper-V and Server Virtualisation

Jeff Woolsey, Senior Program Manager, Microsoft Virtualization
http://blogs.technet.com/virtualization/

Virtualisation is one of the hottest topics in IT today and security is a top priory for IT staff. In this session we cover security best practices for Hyper-V and introduce the Hyper-V Security Guide. This guide is Microsoft’s reference for hardening servers running Windows Server 2008 with Hyper-V enabled.

- Use of BitLocker
- AV scanning of offline VHD images… stale/dormant VMs that get reintroduced to the corporate network then can wreak havoc to new vulnerability exploits.  first product to do this is McAfee VirusScan Enterprise for Offline Virtual Images (rolls off the toungue!)
- AV configuration… added benefit of passthrough disks – host AV will scan these disks.  install AV on the guests!
- VHD performance, 2nd most popular VM workload in Microsoft is SQL! fixed disk performance is now on par with raw/real disk!  but remember spindles still count!
- Dynamic VHDs are now up to 15x faster with R2 – still a 10-15% performance hit over fixed, and the risk of disk over commit
- Multipath I/O (MPIO) in R2 & Win7 is soo much easier with iSCSI Quick Connect
- Advanced Storage Capabilities… storage dedupe and replication, if it is block based it will work
Hyper-V Networking – don’t forget the parent is a VM too!  the Hypervisor slides in beneath the O/S once enabled.  More NICs the better, min 2, min 3 with iSCSI
Jumbo Frames… Significant performance increases, but the infrastructure must support it.  also needs to be end-to-end.  confirm test with ping host -l 4000 -f – if you get a response you have jumbo frames
Virtual Machine Queues - Hyper-V R2 supports processing offloading to newer network adaptors (Intel, Broadcom, etc.), most benefit with 10Gb/E
- more tips… turn off screen savers in guests, in Windows Server 2003 create using 2-way to ensure MP HAL

more of a best practise and walkthrough of some basic tasks like not forgetting to install Intergration Component, good session and great blogger (not me – Jeff! )

Continuing the post on TechEd Europe started on Monday I’ve decided to write daily, there is just too much for one mega-post.  In fact, most sessions have enough quality content to justify their own posts, but I’ll save you from this!    It’s a massive geekfest!  ja!

First session of the day for me is:

CLI302 – How Windows Storage Is Changing: Everything’s Going VHD!

Mark Minasi, internationally recognised technical guru, funny guy, great speaker (his next session on IPv6 in the afternoon was one of the most over-subscribed sessions of the event and I didn’t get in )

Load Windows 7 or Windows Server 2008 R2 on a system, and you’ll notice something sort of strange: there’s no boot record or BCD folder. Look at other Windows 7/R2 systems, and you may notice something even stranger: there’s only one file on the hard disk, and yet you can boot the system and run a normal Windows system. What’s going on here? Simple: Windows 7 gets a lot of press for its faster-than-Vista performance and newer user interface, but there’s a lot more to it also, including native support of VHD files (that’s how a one-file system boots) as well as a new default disk structure, support of direct-to-disk ISO burning, and more. Whether you’re going to Windows 7 sometime soon or five years from now, you’ll want to be prepared for the changes that Windows 7 brings to storage—and who better to prepare you than veteran Windows explainer Mark Minasi? Join Mark for this quick look at Windows 7/R2 storage and save yourself having to read a small mountain of whitepapers!

Mark Minasi: “come to the command line, the command line will set you free!.  If we want the full power my friends… the command line!”

Also, ignore the V in VHD, there’s nothing Virtual about VHD’s

How easy is it to create a VHD?

diskpart
  create vdisk file=filename maximum=sizeinmegabytes [type=fixed|expandable]
  select vdisk=drive:path
  attach vdisk  (can attach as read-only)

Of course, it can also be done through the GUI, although this is lame!  come to the command line!  (I like this guy lots!)

Creating a bootable VHD image:
  1) Take your gold image machine and sysprep it
  2) Attach an extra drive
  3) Image the system to the drive  (/verify /check)
  4) Create a new vhd
  5) Select the vhd
  6) Apply the wim file to the vhd (imagex /apply)
  7) Detach the vhd

You now have a sysprep’d gold image that you can deploy and boot from on as many machines as you like.  Sysinternals also have a tool that does the same and used the same API’s so is functionally identical

Basic VHD boot:
  1) Copy the VHD to a local hard drive
  2) Use BCDEDIT to create a new OS entry that points at the VHD

bcdedit /copy (copy the resultant GUID to the clipboard!)

What about making a system that has no installed O/S – just a pure boot from VHD

If you do this – insomnia!  you cannot hybernate a VHD system!

How to setup WinPE – newsletter #59 on www.minasi.com
Create the system.wim – newsletter #61

MGT213 – Introducing Microsoft System Center Data Protection Manager v3 Beta

Jason Buffington, Technical Product Manager, DPM
http://jasonbuffington.com/
http://blogs.technet.com/DPM
dpminfo@microsoft.com

The third generation of Microsoft’s backup and recovery solution began public beta in October 2009. In this session, we spend most of the hour demonstrating what’s changed, the new capabilities, and what you should be planning for with Data Protection Manager.

Agent deployment changes, can install as before, but also attach to an existing agent which covers for agents behind firewalls, externally installed, etc.

Disk Allocation – changes to support collocating data in the  DPM storage pool, also the ability to Automatically grow the volumes – woohoo!!  (single biggest request for DPM into Microsoft)

Client support – as in client workstations, laptops, etc.  Much improved support for remote clients over long, thin links.  Capabilities to enforce an IT policy for backup but also allow the end user to add to the protection (with controls, i.e. IT can still enforce exclusions such as no MP3s).

“I like one throat to choke” – i.e. using a Microsoft backup tool to backup Microsoft

Using DPM to provide Disaster Recovery, i.e. replace a.n.other replication technologies by having another offsite DPM server (can do very funky automated recovery into a VM using VMM).

One thing to remember if you don’t attend another DPM session.  For compliance, PCI, Sarbanes, whatever!  If you have your offsite DPM server with the tape drive at that location you are already backing up and storing in a seperate geographic location – “You don’t have to pay the guy to lose your tapes for you!”

DPM 2010 Scalability
100 servers, 1000 clients, up to 2000 databases per server
Significantly increases fan-in of data sources per DPM server

DPM Recovery – specify ‘latest’ will not just restore the last DPM replica point, but also play back the transaction logs to the point of failure

www.inframon.com – “scary smart people who have taken DPM management to a whole new level”

you heard it hear 1st…
DPM 2010 from RC onwards will support backup of non-domain joined clients, i.e. none AD clients – this got a major applause (fortunately no ‘high-fives’).

Feedback – Jason is an excellent, engaging presenter with deep understanding of the product with massively relevant content, way more interesting than the SQL performance tuning I was going to do.

BOF04 – Snapshots in Windows Server 2008 Hyper-V

Marcos Nogueira
Snapshots in Hyper-V are a very powerful feature that when used properly can save you from disaster. On the other hand, snapshots, if used incorrectly, can bring a server down. Can snapshot replace a backup? At this BoF session we discuss the best practices for snapshots and some common usage scenarios.

Interactive, audience led session.  Not helped by folks not that keen on talking, and the speaker, sorry host (is not a demo, presentation), didn’t sound confident although should do with 3000 VM’s in his enviornment!

My question about snapshot avhd merge issues, even if you delete the snapshot striaght after the test the avhd/s will continue to grown until either the VM is shutdown or saved.  Is there any other way of merging the AVHD files back in e.g. a ‘live merge’

Using snapshots in development through test, qa, etc. to manage iteration testing

MGT314 – How to Protect SharePoint and Other Application Servers with Microsoft System Center Data Protection Manager

Another session with Jason Buffington of DPM fame, we are now the DPM’ers!

In this session we look at the challenges and special requirements for protecting and recovering Microsoft SharePoint Server, as well as Microsoft Exchange Server, and Microsoft SQL Server. We discuss how Data Protection Manager (DPM) uses the native application VSS writers to ensure supportable backups and recoveries, and discuss the implementation specifics for protecting these key platforms. The discussion focuses mainly on DPM 2007 SP1 implementations for Microsoft Office SharePoint Server 2007, SQL Server 2008, and Exchange Server 2007, but concludes with a glimpse into what DPM v3 is planning for O14 and E14.

DPM2010 new features for SharePoint – auto protection of new content dB’s, no need for a recovery sharepoint farm

Tip – for workload specific DPM information goto www.microsoft.com/dpm/yourworkload e.g. www.microsoft.com/dpm/sharepoint - how to’s, videos, sandboxed demo enviornments, why you need it (for the boss), etc.

Tip – DPM can restore to tape, e.g. when asked for a copy of last month’s backup for the auditor, rather than give them a tape containing EVERYTHING (and you know you’ll never get it back), you can just give them the data they need from the time they want it.

Further training… TechNet virtual labs – for hands-on learning with DPM 2007 SP1

SVR308 – Storage and Hyper-V: The Choices You Can Make and the Things You Need to Know

 
Storage and virtualisation are two of the hottest technologies in IT right now—and they’re even better together. Microsoft is delivering new products into both spaces, and understanding the impacts that storage has on a server virtualisation design is key to understanding and successfully building a virtual infrastructure. This session talks about where we are, where we’ve been, and where we are going. We give you the information you need to know to make the right storage decisions in this new virtual world.

Hyper-V Architecture overview
Key to have the HCL/supported hardware (we know this, but it’s especially important for clustering)

Hyper-V Storage
DAS: SCSI, SATA, eSATA, USB, Firewire
SAN: iSCSI, Fibre Channel, SAS
NAS is not supported (stick an iSCSI target on it!)

Storage parameters / constraints..
Max VHD size 2048GB / 2TB

Tip – ISO”s on network shares, and especially with remote management.. needs a computer account & constrained delegation!  (gets round issue of needing local ISO’s on the nodes… ahhh!  learnt something! )

Hyper-V Constrained Delegation

Differencing VHDs – performance vs. chain length – above 8 the performance delta becomes significant

Hyper-V and AV..

SAN Boot and Hyper-V
  Boot host from SAN (FC or iSCSI)
  Boot child VM frin iSCSI (PXE solution such as emmboot)

iSCSI performance best practises with Hyper-V..
  Jumbo Frames since R2
  ..

Live Migration/HA Best Practises
  CSV
  Multi-Path IO (MPIO) is your friend
  Standardise names of virtual switches
  Use ISO’s not real CD/DVD as will break live migration

Sizing Storage for Hyper-V
  see pic

HA with Hyper-V using MPIO & FC SAN
  see pic

DAT306 – Microsoft SQL Server 2008 Reporting Services Best Practices

Chris Baldwin (chris.baldwin@microsoft.com)

Chris started by admitting to talking far too quickly, so excuse this post, maybe disjointed!

Is there anything you could be doing better with SQL Server Reporting Services? This session explores some of the best practices associated with report design and report server deployment. Topics range from the best way to plan your scale out architecture, to the best way to optimise your reports design for performance.

Agenda..

Server Deployment
  Backup/Restore… see pic
  Security – use read-only accounts, be wary of unecessary use of Windows Integrated, if really paranoid can be disabled for RS (server properties)
  Monitoring & Planning – use VS2005 to perform load testing on RS, RS caches so beware testing results, can use ‘NullRenderer’ subscription to pre-process and prime the cache ahead of use
  One-Box deployments useful, but not great for production due to lack of scale

Report Design and layout
  Understanding Tablix
  Report performance analysis and optimisation – subreports can negatively impac, move subreport layout up into main report, merge main report and subreport datasets

Data Visualisation
  .. will post pics!

RS Architecture..

pics to post!