Published on 15 Sep 2009
In the last posts of this series I gave an overview why performance monitoring is important and that it is not a trivial subject, that terminal servers are an entirely different matter and they require special attention and, in the last post, how to monitor the processor and related corners of the operating system. Continuing my way through the operating system, I’d like to take an extensive look at the memory subsystem in this post.
Published on 26 Aug 2009
As Windows Server 2008 R2 is now RTM and is only available as x64 edition, you need to ask yourself how Windows x64 effects application delivery because sooner or later you will succumb. Although a general adoption of x64 is still a long way gone, many companies are beginning to actively pursue evaluating Windows x64 and testing applications on it. But considering Windows Server 2008 x64 on a very high level, there are pitfalls you need to be aware of.
Published on 06 Aug 2009
After the introductory articles about the importance of performance monitoring and the characteristics of terminal servers in that area, I finally dive into the first and most obvious subsystem of an operating system – the processor. The processor or rather the processor cores are the workhorses of the system. They are collectively responsible for executing commands for each of the active processes. But they also switch the context between running processes to reflect the priorities assigned to each process by the operating system. In this article, I will introduce the well-known as well as some rarely used metrics to monitor the load put on the processor subsystem.
Published on 29 Jul 2009
The Microsoft Remote Desktop Services Team has release a very intriguing article about Aero Glass Remoting with Windows Server 2008 R2. Being a tech guy, I have tested this on a development system and I must say that I am officially impressed.
In the last years, I joined the ranks of those migrating to Windows Vista and, later, to Windows 7 RC not only because it was the next incarnation of the Windows operating system but due to Aero which is part of the Home Premium (or higher) editions.
But soon after the initial euphoria subsided, I asked myself who would actually profit from Aero Glass with or without remoting it? Consumers apparently are but this is not my business.
Published on 29 Jul 2009
In the first article of this series, I provided a short overview why performance monitoring is important, what subsystems are to be monitored and named some tools focussed on monitoring terminal servers. Having been concerned with the performance analysis of terminal servers in many projects, I can draw some conclusions about terminal servers before diving deeper into the subject. I’d like to introduce two categories of terminal servers from a performance standpoint.
Published on 27 Jul 2009
After having published the XmlServiceReader, I have described how to use this tool to customize health check in XenApp Health Monitoring and Recovery (HMR). In this article I will cover health checks that to not apply to a single server but assure the operation of the farm as a service independently of individual servers.
Published on 27 Jul 2009
In my experience, terminal servers are not properly monitored resulting in administrators not knowing how a farm performs – neither concerning the peak performance nor the trend of the handled load. This leads to an inaccurate and often inadequate sizing of the terminal server environment because only rough estimates arise from such a negligence.
In this series of articles, I’d like to expand on the topic and stressing why monitoring is important for all environment (including terminal servers), what needs to be monitored and how is can be achieved.
Published on 20 Jul 2009
When EdgeSight is set up correctly, all data is collected without any user interference especially no administrator credentials are required. Unfortunately, this is only true for historical reports generated from the EdgeSight database. As soon as real-time reports are used and workers are started manually on devices in trusted domains, the administrator’s job gets tricky.
Published on 07 Jul 2009
We all know that it is trendy to use a profile solution to rid Windows of some shortcomings of roaming profiles. And quite a number of you have looked at Citrix Profile Management (also known as User Profile Manager). In its current incarnation, UPM is configured using a group policy specifying the profile path. But similar to utilizing the “Set path for TS Roaming Profiles” for Terminal Services (soon to be Remote Desktop Services), this introduces the limitation that all users logging on to a server receive the same profile path – most most likely with some dynamically substituted components like environment variables or, in the case of UPM, fields from the user object in Active Directory.
Unfortunately, both solutions (UPM and “Set path for TS Roaming Profiles”) are inferior to managing profile paths in Active Directory user objects. The latter enables administrators to distribute users across several servers or use components representing an organisational affiliation. Wouldn’t it be neat to combine those to methods of maintaining profile paths?
Published on 30 Jun 2009
I have recently worked on an implementation for a customer and was concerned with a large number of group memberships. Although the solution for authenticating against the operating system are well documented by Microsoft (here, here and here), the XML service does not allow requests larger than 4KB. So if users have a large number of group memberships, authentication via the XML service can fail due to this limitation. But here’s the solution.