Month: April 2013

The Death Of The Windows Service Pack … Or Is It?

Don’t bother waiting for SP1 before deploying Windows Server 2012.  It’s not coming (to be clear: I haven’t been officially told that or anything else).  It’s looks like Microsoft is switching release cycles from every 3 years to annually.  That makes service packs an impossibility without Microsoft hiring an a lot more developers, testers, and PMs.

What was a service pack?  It was a collection of security and bug fixes.  Security fixes came from the Windows catalog via Windows Update, WSUS, etc.  Bug fixes were normally downloaded manually from the Support site.  A service pack would normally group all these together, maybe add some minor features to avail of recent device enhancements, and be tested by Microsoft as a unit.  This would be a big point release, seen by some as a mark of maturity.

Windows Server 2008 R2 SP1 actually included some significant new features for virtualization: Dynamic Memory for Hyper-V and RemoteFX for Remote Desktop Services.

Since the release of Windows Server 2012 (and Windows 8) we’ve seen update rollups being released pretty regularly via the Windows Catalog (therefore via Windows Update).  These consist of bug fixes, not security fixes.  The security fixes are still coming down as usual.

To be honest, the earlier update rollups irked a few of us because they listed the KB numbers of the bugs they fixed, but those KB articles weren’t publicly available so we didn’t know what bug fixes we were getting.  But it seems like Microsoft listened to our feedback with the release of the April 2013 Update Rollup and listed/linked/summarised the contained bug fixes.

I’ve lost track of how many of these URs have been deployed since RTM.  It could be 3, maybe 4, possibly 5.  They come via Windows Update so I’m pushing them out with the security updates via WSUS in the lab.

So, if you are waiting for “Service Pack 1” for Windows Server 2012, I’m happy to tell you that we’re at least at Windows Server 2012 “SP3” and maybe even at Windows Server 2012 “SP5”.  Stop delaying and join the rest of us in this decade which we’re already 3 years into.

Technorati Tags:

Update Rollup 2 For System Center 2012 SP1 Is Released

Microsoft has released UR2 for System Center 2012 SP1 via Windows Update.  That means you’ll auto download and deploy (pending manual/auto approval on your part) this update via WSUS, etc.  You can also manually download the updates to each product. 

Note that VMM is not included this time around and OpsMgr has quite a few updates.

Please test and then update yours or your customers’ sites to improve the performance and stability of your System Center deployments.  For consultants, this is an opportunity for you do do a little *ahem* sales, and see if there are some further deployments/customisations that you can do for your clients.

App Controller (KB2815569)

  • Issue 1: You cannot change the virtual machine network of deployed virtual machines.
  • Issue 2: The network connection is set to None after you view the network properties of a deployed virtual.
  • Issue 3: You cannot view the virtual networks for a virtual machine.
  • Issue 4: When you change the virtual network in App Controller, you receive the following error message:
  • Issue 5: You cannot copy VMs that have multiple processors or large amounts of memory from VMM to a Windows Azure.
  • Issue 6: App Controller requires Microsoft Silverlight 5 but links to the download page for Silverlight 4.
  • Issue 7: An argument null exception may occur if network connectivity is interrupted.

App Controller Setup (KB2823452)

  • Issue 1: App Controller cannot be installed if the Microsoft SQL Server database server name starts with a number.
  • Issue 2: Setup incorrectly reports that the SQL Server database has insufficient disk space.
  • Issue 3: Setup is unsuccessful when it tries to enable Internet Information Services (IIS).

Operations Manager (KB2826664)

  • Issue 1: The Web Console performance is very poor when a view is opened for the first time.
  • Issue 2: The alert links do not open in the Web Console after Service Pack 1 is applied for Operations Manager.
  • Issue 3: The Distributed Applications (DA) health state is incorrect in Diagram View.
  • Issue 4: The Details Widget does not display data when it is viewed by using the SharePoint webpart.
  • Issue 5: The renaming of the SCOM group in Group View will not work if the user language setting is not "English (United States)."
  • Issue 6: An alert description that includes multibyte UTF-8 characters is not displayed correctly in the Alert Properties view.
  • Issue 7: The Chinese (Taiwan) Web Console displays a wrong message.
  • Issue 8: The APM to IntelliTrace conversion is broken when alerts are generated from dynamic module events
  • Issue 9: Connectivity issues to System Center services are fixed.
  • Issue 10: High CPU problems are experienced in Operations Manager UI.
  • Issue 11: Query processor runs out of internal resources and cannot produce a query plan when you open Dashboard views.
  • Issue 12: Path details are missing for "Objects by Performance."

Operations Manager – UNIX and Linux Monitoring (Management Pack Update) (KB2828653)

  • Issue 1: The Solaris agent could run out of file descriptors when many multi-version file systems (MVFS) are mounted.
  • Issue 2: Logical and physical disks are not discoverable on AIX-based computers when a disk device file is contained in a subdirectory.
  • Issue 3: Rules and monitors that were created by using the UNIX/Linux Shell Command templates do not contain some parameters.
  • Issue 4: Process monitors that were created by the UNIX/Linux Process Monitoring template cannot save in an existing management.
  • Issue 5: The Linux agent cannot install on a CentOS or Oracle Linux host by using FIPS version of OpenSSL 0.9.8.

Service Manager (KB2828618)

  • Issue 1: If the number of "Manual Activities" displayed in the Service Manager Portal exceeds a certain limit, page loads may time out.
  • Issue 2: Incorrect cleanup of a custom related type causes grooming on the EntityChangeLog table to stall.
  • Issue 3: Service requests complete unexpectedly because of a race condition between workflows.
  • Issue 4: The console crashes when you double-click a parent incident link on an extended incident class.
  • Issue 5: PowerShell tasks that were created by using the authoring tool do not run because of an incorrect reference.
  • Issue 6: The Exchange management pack is stuck in a Pending state after management pack synchronization.

Orchestrator (KB2828616)

  • Issue 1: The Monitor SNMP Trap activity publishes incorrect values for strings when a Microsoft SNMP Trap Service connection is used.
  • Issue 2: Inconsistent results when you use Orchestrator to query an Oracle database.

Data Protection Manager (KB2822782)

  • Issue 1: An express full backup job in SC 2012 SP1 may stop responding on a Hyper-V cluster that has 600 or more VMs.
  • Issue 2: When a SC 2012 SP1 item level restore operation is performed on a SharePoint the restore is unsuccessful.
  • Issue 3: When you open DPM on a computer that is running SC 2012 SP1, the Welcome screen does not indicate the correct version of SP1.
  • Issue 4: When you perform a disconnected installation of the DPM 2012 SP1 agent, you receive an error message.
  • Issue 5: When you use DPM 2012 SP1 for tape backup, a checksum error may occur when the WriteMBC workflow is run.
  • Issue 6: Backups of CSV volumes may be unsuccessful with metadata file corruption in DPM 2012 SP1.
  • Issue 7: The DPM console may require more time to open than expected when many client systems are being protected.

Over 1 Year Since The Project Started

Well over a year ago discussions started with the publisher.  A lot of that was “will this version of Hyper-V actually sell?” which leads to “will there be enough demand for a book on this product?”.  6 months of writing by 4 authors and reviewing by another, plus countless hours of editing.  And after all that work, 600 pages just arrived.

You can get 600 pages of Windows Server 2012 knowledge to own from these outlets: Smile

Yes, I know there are some typos still there – please contact Wiley directly.

Frequently Asked Questions: FAQ.

KB2819564–Corruption Occurs When Merging Data Deduplication-Enabled VHDs In WS2012

Strictly a storage (or storage virtualisation) KB article, but relevant to Hyper-V admins, this post describes when data corruption occurs when you merge Data Deduplication-enabled VHDs in Windows Server 2012.

Symptoms

Assume that you enable Data Deduplication on a virtual hard disk (VHD) on a Windows Server 2012-based computer. The computer crashes while you are merging VHD files. In this situation, data corruption occurs in the VHD files.

To resolve this issue, install the Windows 8 and Windows Server 2012 cumulative update 2822241 (available via Windows Update).

KB2829549 – Hyper-V Linux Integration Services & APIC Support In The Guest OS

According to Microsoft in a new fast publish article:

If you are using Linux Integration Services in a Linux Virtual Machine running on Hyper-V Server 2008 or Hyper-V Server 2012, you should enable Advanced Programmable Interrupt Controller (APIC) support in the Virtual Machine.

This APIC setting is configured in the guest OS.  Some Googling reveals that some are disabling this setting using “noapic”.

APIC support is required for proper functioning of all versions of Linux Integration Services. If APIC support is disabled by specifying "noapic" in the Linux Virtual Machine’s boot options, then the Virtual Machine may fail to boot after installation of Linux Integration Services.

Technorati Tags: ,,

KB2821052 – “0x000000D1” Stop Error When Opening MPIO Snap-In WS2012 Computer

This isn’t a Hyper-V bug, but it’s one that those using SAN storage should pay attention to … and that often include Hyper-V admins/engineers: "0x000000D1" Stop error when you open the MPIO snap-in on a computer that is running Windows Server 2012.

Symptoms

Consider the following scenario:

  • You connect a storage device that is controlled by Microsoft Multipath I/O (MPIO) to a computer that is running Windows Server 2012. 
  • You remove all paths to the storage device.
  • You open the MPIO snap-in or try to switch to the MPIO tab on the properties window of a disk that is managed by MPIO.

In this scenario, the computer crashes. Additionally, you receive a Stop error message that resembles the following:

STOP: 0x000000D1 (parameter1, parameter2, parameter3, parameter4)

Notes

  • This Stop error describes a DRIVER_IRQL_NOT_LESS_OR_EQUAL issue.
  • The parameters in this Stop error message vary, depending on the configuration of the computer.
  • Not all "Stop 0x000000D1" errors are caused by this issue.

A supported hotfix is available from Microsoft.

Technorati Tags: ,

KB2823643 – VMs Freeze At “Stopping” State After You Shutting Them Down On WS2012 Hyper-V Cluster

This hotfix is for when virtual machines freeze at the "Stopping" state after you shut down the virtual machines in Windows Server 2012.

Symptoms

Consider the following scenario:

  • You create some failover cluster nodes on computers that are running Windows Server 2012.
  • You have the Hyper-V server role installed on the cluster nodes.
  • You create virtual machines on one cluster node, and you configure the virtual machines as cluster resources.
  • You enable Quality of Service (QOS) on the virtual machines, and you set the QOS to the minimum allowed value.
  • You send lots of network traffic from the virtual machines to another cluster node.
  • After the virtual machines run for a while, you shut down the virtual machines.

In this scenario, one or more virtual machines freeze at the "Stopping" state. Therefore, you cannot perform any actions on the virtual machines. For example, you cannot perform a live migration or a backup.

A supported hotfix is available from Microsoft

KB2823958 – Virtual Switch Extension Can’t Send Packets Separately Through Each NIC In WS2012 Hyper-V

A lot of posts appeared last night that WS2012 Hyper-V admins might want to pay attention to.  This one deals with a situation where virtual switch extension cannot send packets separately through every individual network adapter in Windows Server 2012.

Symptoms

Consider the following scenario:

  • You have a virtual switch extension installed on a Windows Server 2012-based Hyper-V server.
  • The extension sends packets that contain one or more of the following flags: 

    NDIS_SEND_FLAGS_SINGLE_QUEUE
    NDIS_SEND_FLAGS_SWITCH_DESTINATION_GROUP
    NDIS_SEND_FLAGS_SWITCH_SINGLE_SOURCE
    NDIS_SEND_COMPLETE_FLAGS_SINGLE_QUEUE
    NDIS_SEND_COMPLETE_FLAGS_SWITCH_SINGLE_SOURCE
    NDIS_RECEIVE_FLAGS_SINGLE_VLAN
    NDIS_RECEIVE_FLAGS_SINGLE_QUEUE
    NDIS_RECEIVE_FLAGS_SINGLE_ETHER_TYPE
    NDIS_RECEIVE_FLAGS_SWITCH_DESTINATION_GROUP
    NDIS_RECEIVE_FLAGS_SWITCH_SINGLE_SOURCE
    NDIS_RETURN_FLAGS_SINGLE_QUEUE
    NDIS_RETURN_FLAGS_SWITCH_SINGLE_SOURCE

In this scenario, the packets are sent by using the same network adapter in the virtual network.

Cause

This issue occurs because Network Driver Interface Specification (NDIS) chains the Net Buffer Lists (NBLs) incorrectly.

A supported hotfix is available from Microsoft.

Windows Server 2012 Hyper-V Installation And Configuration Guide Is Shipping In Europe

Good news; the paperback version of the new Hyper-V book is now shipping from Amazon in Europe.  Act fast though; The German store reports this morning that there are just 7 left in stock and the UK store has just 11. The book is also available in Kindle (contact Wiley about other e-versions) and was, of course, already available in the USA store.  I haven’t looked at other regions yet.

9781118486498 cover.indd

Hyper-V Recovery Manager – Orchestration of Hyper-V Replica Failover

Currently in limited preview, Hyper-V Recovery Manager (a part of Windows Azure Recovery Services) provides orchestration of Hyper-V Replica replicated System Center managed clouds.  The concept is:

  • You have a System Center managed cloud in site A.
  • You use Hyper-V Recovery Manager to orchestrate replication via Hyper-V Replica to site B
  • Hyper-V Recovery Manager is used to coordinate failover.

To participate in the limited preview, you must have a Windows Azure account.  Candidates from the program must be from a small set of countries: United States, Canada, United Kingdom, Germany, France, Belgium, Switzerland, Denmark, Netherlands, Finland, Australia, Japan, India, or New Zealand.  Well that rules me out then!