While constantly checking Task Manager and Resource Manager, I would sometimes notice high CPU utilization due to what I call the 3-headed monster. Here’s the definition for that:

3-headed monster

The user experience of Windows Indexing, System Restore, and/or Windows Error Reporting causing less than desirable states of desktop responsiveness especially during logon process. These are three Windows “features” I would consider disabling to improve performance immediately following logon. Disabling these will also improve overall performance.

Indexing

Let’s start with indexing. The theory of indexing is sound and I think can be applicable to home desktop users. However; on workstations in the business environment, I’m not so sure.

I’ve seen indexing cause slow performance when the Users directory is included in indexing. If you read my previous post, you learned how to reduce the size of the user’s roaming profile. That is helpful for the profile synchronization process but doesn’t reduce the size of the local profile. This local profile can be very large and could contain a very large number of files. I’ve found that removing this directory from indexing can greatly improve performance.

Here’s how to disable it.

From the Control Panel search for Indexing. Click Change how Windows searches. The Indexing Options dialog box will display showing you the number of items indexed. Click Modify to disable Indexed Locations.

Indexing Options and Indexed Locations

System Restore

Now on to System Restore. This is another feature that I feel has a better place in the home environment. It allows you to recover easily should something go wrong. But again, I have seen this cause massive system lagging when it’s running. I recommend disabling it and here’s why. Most shops have at least one spare machine lying around. Keep that as a hot spare and replace that machine that took a nose dive into the concrete-filled swimming pool. Roaming profiles and server data storage mean that in this case that machine is just a dumb terminal anyways. Turn off the extra services in favor of performance for your users and they’ll be much happier.

Here’s how to disable it.

Right-click Computer and select Properties. Click Advanced system settings and then the System Protection tab. Select a drive and then click Configure. Select radio option to Turn off system protection.

Disable System Restore

Windows Error Reporting

Just a quick disclaimer, I’m a big fan of Microsoft. They’re aren’t perfect but they sell good products that are universally recognized by anyone. In fact, I think they’re so great that I’m willing to bet they don’t really need every one of my workstations reporting every little problem. As with System Restore, I’m interested in removing anything that is going to cause lagginness (if that’s a word) for my end user.

Very rarely, I’ve seen the WER directory grow very large. Disabling indexing on the Users directory should mean that a large WER directory isn’t too big of a deal. The location of the WER directory is typically C:\Users\%username%\AppData\Local\Microsoft\Windows\WER.

Here’s how to disable it.

Windows Error Reporting can be disabled using group policy. If you want to disable by computer, create a GPO and configure Computer Configuration\Administrative Templates\System\Internet Communication Management\Internet Communication settings. To disable by user, customize the GPO section User Configuration\Administrative Templates\System\Internet Communication Management\Internet Communication settings.

Turn off Windows Error Reporting

In both the computer and user settings, there are also several other communication settings that I would highly recommend you review and consider disabling.

For additional information please refer to this TechNet article.

A couple other reminders that I may be helpful: Make sure that anti-virus and anti-malware are set to scan at times when user’s are not active. Also make sure that server backup activity is not affecting user performance.

Also read 10.5 ways how to speed up a slow Windows

• Troubleshoot slow logon – Part 1: Profile size (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• FREE: Smart Defrag – A defragmentation tool (9)
• FREE: NovaBench – Free benchmark software (1)
• Turn off indexing in Windows 7 and Windows Vista? (5)

I recently spent some time troubleshooting slow logons. I had several complaints and had also personally witnessed some very slow logons. After some investigation, I found that nearly all of the user’s had either a very large profile or had directories with extremely high file count. Rather than just delete and hope the problem didn’t happen again, I decided to seek out a solution that would prevent this from happening in the future.

As you may know, just before discovering a solution, there’s usually a key discovery that lowers your blood pressure and allows your confidence level to return to normal. For me, that moment was discovering the power of using the Exclude directories in roaming profile and Limit Profile size Group Policy settings together.

Exclude directories in roaming profile and limit profile size

The Group Policy Object (GPO) setting Exclude directories in roaming profile is one of those settings that is unusually self explanatory. It allows you to create a list of directories, from the root of the user roaming profile location, that should be excluded from roaming profile synchronization.

How do I configure it?

This group policy setting is located in the User Configuration/Policies/Administrative Templates/System/User Profiles section of a GPO. To add multiple folders, separate them in the list using a semi-colon. For example, to exclude the Cookies directory and the Mozilla directory use this for your value:

AppData\Roaming\Microsoft\Windows\Cookies;AppData\Roaming\Mozilla

Exclude directories in roaming profile

What to exclude

Sometimes profiles grow large and you may not be sure which directories are the cause. This is where the Limit Profile size setting comes in very handy. Enabling the setting allows you to specify what the Max profile size should be as well as the message to use to notify the end user. This could be a double-edged sword. It can prompt your user to contact you when their profile is large but should be used with caution depending upon your user group. In my experience, I found this was a great way to target those experiencing slow logons due to profile size.

For the admin, the real power in this setting is the Profile Storage Space dialog box that is enabled in the system tray of the user’s desktop. It allows you to quickly view a list of the files in the profile sorted by size. I found this to be invaluable. That is because the cause of each user’s large profile can be unique.

Profile Storage Space dialog box enabled with Limit profile size

In your environment, I would recommend enabling this setting with a large Maximum profile size of somewhere around 300MB for those user’s that you know are having problems with slow logon. This will allow you to begin to assess the cause of the large profiles. As you add directories to the Exclude directories in roaming profile setting, you will see directories disappear from the Profile Storage Space view.

• Modify Exclude directories in roaming profile GPO setting that applies to the user
• Synchronize domain controllers
• Refresh group policy by running gpupdate /force from the command line
• Log the user off and then back on
• Double click the Profile Storage Space icon in the system tray

Ideally, you want to get profile size below 50MB. Below is the list I’ve compiled that has taken care of this issue for most of my users. This should work for Windows Vista and later desktop users. AppData\Roaming\Microsoft\Windows\PrivacIE;AppData\Roaming\Microsoft\Windows\Cookies; AppData\Roaming\Mozilla;AppData\Roaming\Adobe;AppData\Roaming\Macromedia;$Recycle.Bin; AppData\Local;AppData\LocalLow

In the next article, I’ll discuss how customizing the three-headed monster (Indexing, System Restore, and Windows Error Reporting) can improve user logon and desktop performance for domain users.

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• FREE: Smart Defrag – A defragmentation tool (9)
• FREE: UserProfilesView – View user profiles (0)
• FREE: NovaBench – Free benchmark software (1)

Some IT departments are under governmental and/or industry regulations that require them to account for all system downtime. Other IT departments may be mandated by their service-level agreements (SLAs) to provide documentation for any server restarts or shutdowns.

In Microsoft Windows Server 2008 R2, we have the Shutdown Event Tracker to assist us in not only the previously mentioned scenarios, but also for any shutdown-related troubleshooting we may be called to undertake.

How the Shutdown Event Tracker works

In a nutshell, the Shutdown Event Tracker enables Windows systems administrators to provide a reason for any system shutdown or restart event. Let’s clear up some terminology before we proceed any further:

• Planned: This is a shutdown or restart event that was accounted for in advance
• Unplanned: This is a shutdown or restart event that was not accounted for in advance
• Planned + Unexpected: This is a planned shutdown or restart event (for instance, using the power button to shut off a server instead of the Shut Down command) that was unexpected by the operating system
• Unplanned + Unexpected: This is an unplanned shutdown or restart event (for example, a power failure) that was unexpected by the operating system

When you perform an orderly shutdown or restart in Windows Server 2008 R2, the default behavior of the operating system is to display the Shutdown Event Tracker dialog box. This is shown in the next figure:

Shutdown Event Tracker dialog

The systems administrator has an opportunity to (a) specify whether the event is planned or unplanned; (b) choose an event category; and (c) provide a comment.

Shutdown-related events are stored in the System event log under Event ID 1074. You should periodically filter this log to isolate and review shutdown-related events.

Shutdown Event Tracker events in Event Log

If your server crashes or suffers an otherwise unexpected restart, the Shutdown Event Tracker is invoked automatically upon the first administrator logon to the system.

How to Disable the Shutdown Event Tracker

If you want to disable the Shutdown Event Tracker, you can do so either by modifying the local Group Policy of the target system, or by editing the Windows Registry.

To use the Group Policy method, open the Run dialog box from the start menu and type gpedit.msc to open the machine’s local Group Policy.

NOTE: You can manage the Shutdown Event Tracker for multiple systems through Active Directory Group Policy.

In Group Policy Editor, navigate to Computer Configuration\Administrative Templates\System, open the Display Shutdown Event Tracker policy, and set it to Disabled. This process is shown in the following screen capture:

Disable Shutdown Event Tracker through Group Policy

If you prefer to use the Registry to perform this action, then type regedit from the Run dialog or an administrative command prompt and navigate to HKEY_LOCAL_MACHINE\SOFTWARE\Policies\Microsoft\Windows NT\Reliability. Next, open the ShutdownReasonOn value and set it to 0.

Disable Shutdown Event Tracker through Group Policy

Conclusion

By now you should understand the purpose and basic functionality of the Shutdown Event Tracker in Windows Server 2008 R2. You also should know how to disable the feature by using either Group Policy or the Windows Registry.

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• FREE: Smart Defrag – A defragmentation tool (9)
• FREE: NovaBench – Free benchmark software (1)
• Turn off indexing in Windows 7 and Windows Vista? (5)

The two defragmentation tools, I reviewed previously (MyDefrag, Disk Defrag), have good features for manually controlling the defragmentation process. By contrast, IObit Smart Defrag certainly lives up to its name because automatic defragmentation is its strength.

Most defragmentation tools support scheduling, and so does Smart Defrag. However, this feature is disabled by default. This makes sense because Smart Defrag has something better to offer: Automatic Defrag.

The problem with scheduled defragmentation is that it can’t always guarantee that the computer isn’t busy at the scheduled time, and a desktop or laptop could be powered off at the scheduled time. Automatic Defrag always runs in the background and only starts defragmenting if the PC is idle. You can configure the PC’s idle time, after which Smart Defrag becomes active, and you can tell the tool to pause whenever the resource usage exceeds a configurable percentage.

Smart Defrag’s Boot Time defrag is another function that most free defragmentation tools lack. Some system files can’t be defragmented during runtime because this would make the system unstable. You can configure Smart Defrag to defragment the pagefile, the hibernation file, the Master File Table (MFT), and other problematic system files when the computer boots up.

Yet another “smart feature” is quite useful for laptop users. Defragmentation can be automatically stopped when on battery. As disk activity is one of the biggest power consumers in a laptop, this feature will certainly improve the battery life noticeably.

You can’t manually defragment specific files and folders as with Auslogics Disk Defrag. However, you can exclude files and folders from automatic or scheduled defragmentation. This can be useful if you don’t want the tool to mess with an important database file.

Smart Defrag can also move frequently used files and directories into the fastest area of the disk. Unfortunately, it is unclear how the tool determines what files are frequently used. I guess, this feature only works for system files.

All in all, Smart Defrag is my favorite defragmentation tool from the three tools I have tested. It has the nicest user interface, and the background defragmentation convinced me. For scripting guys, however, MyDefrag is the best choice; if you often have to defragment specific files, Auslogics Disk Defrag is your best bet.

I tested Smart Defrag v2.0.1 on Windows 7.

IObit Smart Defrag

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• FREE: Disk Defrag – A disk defragmenter (3)
• FREE: MyDefrag – Script defragmentation (4)

The free benchmark software NovaBench is a simple tool, but it offers all essential benchmark tests. Benchmark tools that are not free usually have more sophisticated benchmarking functions with nice graphical outputs. NovaBench’s result page is rather simple, giving you just the plain numbers of the benchmark tests. However, if you just want to know how fast the components in a PC are, this is all you need.

NovaBench tests the RAM speed in MB/s, tests the CPU speed (floating point, integer, MD5 hash generation), tests graphics performance by measuring the 3D frames rate, and tests write speed of the hard disk.

NovaBench calculates a proprietary score that allows you to compare the performance of different PCs. If you register at the NovaBench website, you can store your results online.

Unfortunately, NovaBench has to be installed. A benchmark software tool is much more useful if it is portable. I copied the programs folder to another computer to see if NovaBench would work without installation, but the benchmark stopped when the RAM test was running. I guess a DLL was missing in the installation folder.

I tested NovaBench 3.0.3 on Windows 7.

Another nice free benchmark software is the hardware information tool HWiNFO32. Also check out the 4sysops list of free benchmark tools. What is your favorite benchmark software?

NovaBench

• Visualizing your server rack with Microsoft Visio 2010 (10)
• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• FREE: Smart Defrag – A defragmentation tool (9)

One of the reasons why many have experienced Windows Vista to be slow was because the Windows Search indexing process kept the hard drive busy. A hard disk that is always busy can slow down even a well-equipped PC considerably. This is why I recommended to disable indexing in Vista at that time. The question now is whether it also makes sense to turn off indexing in Windows 7.

I don’t know what Microsoft did in Windows 7 with Windows Search, but it appears that improvements were made in the detection of active applications that could be impaired by the indexing process. I never saw a Windows 7 computer being slowed down by the indexing process. The latest Windows indexing technology is also available for Vista and Windows XP in Windows Search 4.0, which may eliminate all performance issues related to indexing.

I wrote this article because my previous post about disabling Vista indexing is still read about 4000 times per month. Thus, it seems that many people still believe that turning off indexing improves performance. This might be because many don’t know about the update for Windows Search. It could also be that they are running Vista on old hardware that is not suitable for this modern operating system.

However, I do believe that even an old computer running Windows Search 4.0 won’t perform much better with indexing turned off once initial indexing has been completed. The only reason I can think of that would justify disabling Windows Search is because you want to run a third-party desktop search solution.

The other question is whether there really is a better desktop search solution for Windows. Windows Search is perfectly integrated into Windows, does not require a web browser, and works well together with Outlook/Exchange and other Microsoft enterprise search products.

But if you still want to turn off indexing in Windows 7, you have several options. I recommend having a look at the comprehensive Windows Search FAQ first, so you know what you are actually turning off.

Options to turn off indexing in Windows 7

You can remove some or all folders that are indexed in the Indexing Options applet (type “indexing” at Start Search).

You can also just pause indexing for 15 minutes if you want to be sure indexing doesn’t interfere with another hard disk intensive task.

You can remove individual drives or folders from indexing through the properties dialog in Windows Explorer.

The most reliable way, certainly, is to disable the Windows Search service (type “services” at Start Search).

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• FREE: Smart Defrag – A defragmentation tool (9)
• FREE: NovaBench – Free benchmark software (1)

Did you ever need quick access to your computer but the wait for boot process completion seemed like eternity? These are the moments when many come to the conclusion that Windows is a slow operating system. However, it is a well-confirmed psychological fact that the human brain has its own subjective timeline, which often has little to do with objective time measurement.

For instance, it is beyond the capabilities of my brain to measure browser speed differences, which, according to various benchmark tests, must be huge. But when it comes to boot speeds, my brain tends to measure much bigger boot-up times than objective measurement tools do.

This also applies to BootRacer. The boot speed this free tool measures doesn’t confirm my own experience. However, a stopwatch more or less verified BootRacer’s calculations. This free boot up speed measuring tool calculates the “time to logon” (not counting BIOS start-up time) and the “time from logon to desktop.” Both time periods added result in the “total time to desktop.”

You might wonder why you need a tool to measure the boot speed if you can get the same results with your watch. First of all, BootRacer is easy to use. You only have to click “Start” and wait until the PC reboots. Second, if you always use the same tool to measure the boot time, you can easily compare different PCs. And third, BootRacer is able to store the history of boot times. Thus, if you are tuning a PC to speed up boot time, you can very easily verify how successful boot speed optimizing was.

BootRacer can also write the boot time to the event log. So if users complain about slow boot speed, you can centrally collect this data with an event log management tool. BootRacer can be configured to measure the boot time every time the PC boots up. I tried BootRacer 3.0.3.

BootRacer

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• FREE: Smart Defrag – A defragmentation tool (9)
• FREE: NovaBench – Free benchmark software (1)

Seasoned sys admins have a good idea of how to use Task Manager and Performance Monitor to understand what’s going on when there’s a server performance issue. Those skills transfer well to the virtual world but there are some gotchas that we’ll cover in this article.

Task Manager in VMs lies

The first rule is don’t ever measure performance of a VM from within a VM. Most sys admins first reaction to performance complaints will be to have a look in Task Manager. Unfortunately that doesn’t work in a VM because it can only see its little keyhole view of the world. Here’s a simple test to understand this; run one VM on a host, run an application (such as Prime95) that maxes out the virtual CPU(s). Task Manager will report 100% CPU but if you now start another VM and run the same application in that VM, Task Manager will still report 100% but it’s actually running at half the speed compared to earlier. This example disregards additional processors in the host but clearly demonstrates: Task Manager in VMs lies.

Create baselines

The second rule for performance monitoring, whether physical or virtual is to create baselines when things are humming along. Unless you’ve got a baseline when performance was good and users happy you have nothing to compare with when issues arise. Learn how to create Data Collector Sets; log for a few days and archive the results.

Use Hyper-V specific counters

When using Performance Monitor in the host, be aware that it might lie as well, don’t use normal counters as they often don’t understand VMs, use Hyper-V specific counters. For processor performance use the Hyper-V Hypervisor Logical Processor\% Total Run Time counter; this will let you monitor physical processors in the host. For VMs use Hyper-V Hypervisor Virtual Processor\% Guest Run Time which lets you monitor virtual CPUs for each running VM (or as a total). As a rule of thumb, less than 75% for this counter (total for all VMs) is healthy, over 75% is a warning sign and more than 85% should definitely be investigated.

Committed Memory

In today’s Hyper-V world, B DM (Before Dynamic Memory), you should monitor \Memory\Available Mbytes on the host, healthy is more than 10% of RAM free, at less than 10% it’s a warning than something is going on and if it’s less than 100 MB it needs to be looked at. Old hands will know that “High pages per second” means a low memory condition but there are cases where this isn’t true (due to memory mapping of files), for more information see here. Use the \Memory\Committed Bytes counter to see how much memory each VM is using when planning your VM memory allocation.

Average Pressure

In the After Dynamic Memory (with SP1) world watch the \Hyper-V Dynamic Memory Balancer\Average Pressure counter, healthy is less than 80; a value between 80 and 100 deserves attention whilst over 100 indicates a critical condition.

Memory buffer

In the physical world as sys admins we always like to have a bit extra up our sleeve. If you were designing a server that you knew would use about 1.5 GB of memory you’d install 2 GB just to be on the safe side. In the VM world this is adjusted with the buffer, the default is 20%. For workloads that need a lot of file cache increase this value. This screen is also where you can assign a higher or lower weight to each VM which will allow a VM to have a higher (or lower) priority for memory as the host starts running low on available memory.

Disk latency

Monitoring disk is done via the \LogicalDisk(*)\Average Disk Sec\Read or Write which indicates your disk latency. Healthy is less than 10ms (0.010), if it goes up to 15ms or above (0.015) it’s time to check it out, at 25ms or above (0.025) the situation is critical.

Network monitoring

For network monitoring use the counter \Network Interface (*)\OutputQueue Length, less than 1 on average is healthy, warning is when it’s above 1 on average, critical is when it’s 2 or more on average.

Conclusion

In this article series we’ve looked at tips for how to design a Hyper-V environment for best performance, some tricks and gotchas to look out for and finally we covered how to monitor performance in a virtualized world.

Resources

Here are a list of websites and whitepapers that you can use to delve deeper into this fascinating topic.

Virtual Reality Check compares performance of hypervisors.

Download a performance tuning whitepaper for both the physical and virtual world from Microsoft for Windows Server 2008 R2 here.

In-depth information on performance tasks

The Performance Analysis of Logs is a PowerShell script that can help you out in deciding which performance counters to collect and how to analyse them.

For disk performance tasks have a look at Iometer

Detailed technical information on how to measure performance in Hyper-V

• Poll: Are you currently using a monitoring solution? (11)
• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• SCOM 2012 review – Part 8: Dashboards (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)

In the last post of this Hyper-V performance tuning series I discussed memory, storage and networking. Today I will list some well-known and some less well known tips and gotchas that will improve your experience with and performance of Hyper-V.

Integration components

Loading the (latest) integration components (IC) is the single most important step in optimizing performance. Check in Device Manager in the VM that Virtual Machine Bus shows up under System Devices, if it does the IC are loaded.

Guest OS

Use newer operating systems as guest OSs. In the physical world a rule of thumb is “if you’re going to run it on older hardware, run an older OS” which would tend to favour spinning up a Windows Server 2003 box rather than 2008 on a less powerful box. In the VM world the opposite applies, if you’re resource constrained on a host but you need a new VM go for the latest OS; it will work better in the virtual world.

Hyper-V Manager

Close Hyper-V Manager when you’re not using it; the monitoring of statistics cause unnecessary WMI activity in the parent partition and the thumbnails of the screens waste resources in both host and guest. Don’t use Virtual Machine Connection to manage VMs, use remote desktop sessions instead, the resource usage in the VMs will be lower.

Services

Don’t run unnecessary services in the host partition, not only will you affect the performance of VMs but you may also be in an unsupported configuration. According to Microsoft you can only run management and backup agents in the parent partition; no other workloads.

Host OS

In a production environment Server Core or the free Hyper – V server are recommended as the host “OS” as they have less overhead (approximately 80 MB less commit charge); both also seriously limits the possibility of someone running unwanted applications in the parent.

Background CPU activity

Minimize background CPU activity in VMs by removing remove unused devices (COM ports), disable the screen saver and leave the VM at the logon screen when not in use. For client OS virtual machines in a VDI environment it’s best to disable SuperFetch and Windows Search as well as scheduled defragmentation. Remove the CD drive from VMs when possible, not just “no media” but actually remove it, otherwise Windows in the VM checks every one second to see if there’s media in the drive.

Network configuration

Dedicate a 1 Gbps NIC for Live Migration in a cluster and avoid the legacy network adapter for all networks (unless you can’t because you need it for PXE booting or because the OS doesn’t support Integration Components). In VMs the emulated NIC consumes CPU resources; use the Synthetic adapter instead. Additional NICs will be necessary on the host for cluster heartbeat, backup, and management. If you’re using iSCSI look to have two virtual NICs on the host, assigned to different virtual switches / physical network cards to provide redundant paths to the storage. In a cluster your Virtual Network names need to be identical across all hosts or Live Migration / Quick Migration won’t work.

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• Hyper-V upgrade – Conclusion (4)
• Hyper-V upgrade – The process (2)

In part 2 of this series about Hyper-V performance tuning we continue the look at designing your Hyper-V farm and the need to balance budget against resource needs; you’ll want to design with a good mix of memory, disk and networking resources to maximize performance.

Memory

Optimizing memory for VMs is a challenge in Hyper-V of today as the memory you assign to each VM is fixed whether the VM actually uses it or not. The good news it’s going to become a whole lot easier when Microsoft releases SP1 for Windows Server 2008 R2 and Dynamic Memory comes into play. There’s a registry key that you can use to tweak the memory reserve for the host but normally you shouldn’t have to touch it, (it’s located at HKLM:\SOFTWARE\Microsoft\Windows NT\CurrentVersion\Virtualization\MemoryReserve). This key is a REG_DWORD and you have to create it, default decimal value is 32 for 32 MB, maximum is 1024. If you follow best practises and run minimal services in the host the default should work fine.

Microsoft recommends setting start-up RAM for 2008 R2, 2008 , Windows 7 and Vista to 512 MB whereas Windows 2003 and XP should start with 128 MB.

My testing indicates that the issues that sometimes caused problems in earlier versions of Hyper-V where there wasn’t enough memory reserved for the host on a heavily loaded system are solved in SP1.

Storage

Storage is always a tricky part of server design and no less so in the virtual world. The main issue with any virtualization platform (unless you’re using pass through disks) is that any disk optimization that an application does is lost as the “hard drive” that the application sees is actually a huge file stored on a drive. This makes random IO latency increase. An example that comes to mind would be Exchange 2010 databases which are optimized for sequential IO; all that optimization is lost in the world of VHDs. The only way to work around this issue is to increase the speed and decrease the latency of the underlying storage system, the best solution (albeit not the cheapest) is to employ SSDs for VHD storage, especially if you have multiple VHDs on the same drive.

Another design decision is around which type of VHD to use; fixed size or dynamic. The golden rule used to be to use fixed virtual disks as they’re much faster than dynamic disks. The performance gap is closing however and dynamic disks are very close to fixed disks in Hyper –V 2008 R2 as are differencing disks. There will still be some overhead involved when a VHD has to grow, not to mention the possibility of running out of disk space due to oversubscribed storage so sticking with fixed disks in production environments is still best practise.

Networking

A mistake I sometime see is looking mostly at memory and processing power when sizing a Hyper-V platform. Networking is very important, allocating enough network capacity to VMs is vital, especially if your VMs are busy file- or database- servers. If you can afford it 10 GB Ethernet is a fantastic way of allocating enough pipe. If you’re employing iSCSI for VHD storage use Jumbo Frames and unbind unnecessary services such as File Sharing and DNS from the iSCSI network card(s). Another important recommendation for network performance is to make sure that the NICs you use are server class, cheaper NICs can cause many problems. Server class NICs will also enables other new NIC features that are offered in 2008 R2 such as Virtual Machine Queues (VMQ) and TCP Chimney Offload. Enable VMQ only for VMs with heavy inbound network traffic.

In my next post I will list list some Hyper-V performance tuning tips and tricks.

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• Hyper-V upgrade – Conclusion (4)
• Hyper-V upgrade – The process (2)

Hyper – V, Microsoft’s answer to VMWare ESXi / VSphere has now been around for a few years and is (arguably) catching up on the feature front. Whether the lower price point and familiar management tools will win businesses over remains to be seen but one thing is certain; when you’ve shelled out for that (or more likely “those”) beefy host servers you want the best performance you can get.

In this article series we’ll first look at design ideas for a Hyper-V environment, then some performance tips and gotchas and finally recommendations for how to measure performance.

One of the first hurdles you have to overcome when you’re designing your Hyper-V infrastructure is working out how many VMs will fit on each host. Today the main limitation for VM density is memory; this will also be true after the introduction of Dynamic Memory in SP1 for Windows Server 2008 R2. Another issue is how to assign processor power to different VMs.

Virtual processors

When Microsoft’s documentation talks about “logical” processors (LPs) they’re referring to physical cores in a system. Virtual processors (VPs) are of course what VMs see. Different guest operating systems can take advantage of different numbers of virtual processors, Windows 2008 and 2008 R2 can see up to four, Windows Server 2003 can see two, Red Hat Enterprise Linux and SUSE Linux Enterprise can see up to four. Windows 7 can see four, Vista can use two and XP SP3 can see two, for more details see here. Hyper-Threading (HT) in Intel processors makes each core look like two, the performance impact that early versions of HT had are gone but HT doesn’t double your CPU capacity so when calculating your logical to virtual CPU ratio count each core as an LP, not each HT core.

Should you assign one VP to each VM or assign more – perhaps the maximum for the OS? There is some overhead in having multiple VPs in a VM due to cross processor communication but the penalty is less with each version of Windows, hence for 2008 R2 VMs it’s safe to go with 4 VPs whereas you should really do some testing for Windows 2003 VMs under normal workloads to see if they need more than one VP.

According to Microsoft as a general rule of thumb it’s best to have four virtual processors per logical processor in the system, maximum is eight. But the question of course is how can you find out the ratio on your hosts? You could manually check the number of allocated virtual CPUs for each VM and compare that against the number of cores in your physical processor(s), a process that doesn’t sound very palatable. Especially not in a cluster where VMs might move around a lot, changing your ratio all the time. The simple answer is to run this one line PowerShell cmdlet which will give you the answer.

If you’re buying new metal and you have the choice make sure your processors support Second Level Address Translation (SLAT); AMD refers to this as Rapid Virtualization Indexing (RVI) and earlier called it Nested Page Tables (NPT); Intel calls it Extended Page Tables (EPT). Without SLAT each VM will take up an extra 10-30 MB and overall processor utilization will be up by about 10%. SLAT can make a huge difference for some workloads, Remote Desktop Services / Terminal Services for instance can see up to 40% more sessions on a VM with a SLAT host. Also look for processors with large L2 and L3 cache; this will help with processing for VMs with applications that have large working sets.

In the next part of this series I will discuss memory, networking, storage and networking.

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• Hyper-V upgrade – Conclusion (4)
• Hyper-V upgrade – The process (2)

After the download had finished I unzipped the archive, started the program, and immediately got an error (CrystalDiskMark). I didn’t want to waste time figuring out what the problem was, so I just downloaded another tool from the Google result list from my previous search. This application (hdtune) did execute without any problems but after all tests were run an average transfer speed of 1419 Mb/sec was displayed. Quite impressive for a SAN with SATA drives.

So I tried the next one (iometer). After the first benchmark run it seemed that this tool is reliable. But once again I was disappointed. After five runs I had benchmark deviations of 200%.

Then I remembered a tool used by one of the magazines I read called H2bench. This tool should produce reliable results. Though the tool’s homepage is only available in German the tool itself is also translated into English. It is a CLI only tool, but its usage is pretty straightforward. The following command will benchmark hard disk number 0, display its results in English, and write a report in the file ‘report.txt’:

H2bench –english –a 0 –w report.txt

Make sure that you start the tool with administrator privileges, because H2bench accesses the hard disk directly and needs to be run elevated. For this reason it also only measures the read performance. If you want to test the write performance make sure there aren’t any partitions on the drive and use the option ‘-!’.

H2bench

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• FREE: Smart Defrag – A defragmentation tool (9)
• FREE: NovaBench – Free benchmark software (1)

The Outlook-like user interface makes it easy to access all the integrated tools fast. There are five sections: Hardware, Configuration, System Files, Resources, and Benchmark.

The Hardware section has 13 tools: System Summary, Mainboard, Processor, Video, IO Ports, Drives, Printers, Devices, Multimedia Network, Power Status, Voltage/Temperature/Fans, and Windows Hardware Experience Index. Some of the information provided by these tools can also be accessed with the onboard Windows System Info tool, but PC Wizard 2008 presents them in a nicer way and also offers more details.

When minimized, PC Wizard shows a transparent On Screen Display gadget which presents information about the current CPU load and the temperature of each processor core. PC Wizard requires only approximately 16MB RAM, so you can let it run in the background if you want to be informed constantly of how much your computer is sweating under the load you are imposing on it.

The only complaint I have about the hardware detection is that my eSATA adapter wasn’t recognized as a disk controller in the Drives section. Instead PC Wizard listed it as a SCSI controller in the Devices Section. The same applied to the RAID controller. My eSATA hard drives were recognized correctly, though.

The Configuration section offers 21 tools which provide information about any Windows setting you can think of. Just to give you an idea of how detailed the information is here, PC Wizard tells you the number of installed Vista Sidebar gadgets and Web Browsers. What was a bit scary is that it displayed my Live Messenger password in clear text. Now you know why you shouldn’t store passwords in applications.

Not all the information provided in the Configuration section was correct. It claimed that the total memory of my laptop is 4GB even though it was equipped with 8GB of physical memory. This reminded me of my experiences when I bought this notebook recently. All PC shops (except the guys at Dell) claimed that there is no such thing as a laptop with 8GB. Perhaps PC Wizard uses the same information source.

More likely is that PC Wizard had problems with Vista x64 because the information in the Details pane was contradictory. It even wanted to make me believe that my new baby only has 1MB of physical memory. I must admit I was a little offended. Anyway, in PC Wizard’s defense it ought to be said that the Hardware section displayed the correct memory information.

Some of the benchmark tools notified me that the test results will not be significant because I run them on a 64-bit system. All of them worked, though. I was quite pleased about the results, so I tend to believe that the benchmarks were correct. These are the system benchmarks PC Wizard supports:

• Processor (Dhrystone (MIPS), Whetstone (MFLOPS), Mandelbrot fractal…)
• L1, L2, L3 Cache, RAM (Bandwidth, Latency…)
• Main Memory (Bandwidth, Latency…)
• Hard Drives
• CD/DVD Rom
• DirectX 3D
• Video
• Removable/Flash Support
• MP3 Compression
• VISTA Experience Index

The other two sections are rather unspectacular. System Files provides the contents of 21 common system files (boot.ini, system.ini, Eventlog, etc.) and Resources offers information about Interrupt Requests (IRQs), Direct Memory Access (DMA), I/O Ports, Memory Resources, and Network Connections.

PC Wizard 2008 has quite a few features for a free tool. It would be great if there were a portable edition. This would allow the help desk guys to get an overview of the PC’s status quickly. I tested PC Wizard 2008 version 1.871.

What is your favorite system information tool?

PC Wizard 2008

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• FREE: Smart Defrag – A defragmentation tool (9)
• FREE: NovaBench – Free benchmark software (1)

Process Lasso is the perfect name for a tool that brings a mad program to its senses. The program’s ProBalance feature constantly monitors all processes, and if an application’s claim for CPU resources exceeds a configurable threshold, Process Lasso will reduce the priority of that program’s process. As soon as the process’ CPU usage falls below a second threshold, Process Lasso will restore the original process priority.

To understand how Process Lasso works, you can use Cpukiller, the tool I blogged about yesterday, as a demonstration. Cpukiller runs at high priority and as a foreground process. Thus, you have to allow Process Lasso to manage foreground process, and processes with non-normal priority, and (see screenshot). If you then let Cpukiller consume 100% of your CPU’s resources, Process Lasso will automatically set Cpukiller’s CPU priority to “Below Normal.”

Even though Cpukiller is still running, your PC’s responsiveness will be normal. If you then reduce Cpukiller’s slow-down factor to below Process Lasso’s threshold, Cpukiller’s process priority will be automatically reset to “High.” You might also like to try running Cpukiller at a slow-down factor of 100%, without letting Process Lasso interfere. The responsiveness of your PC will then be close to 0.

Process Lasso has a couple of other features that can be helpful in improving the responsiveness of a computer. You can configure default process priorities, default CPU affinities, boost foreground processes, configure disallowed processes, limit the number of instances of a program and more.

Process Lasso is only free for private and academic use. The prices for the professional version start at $24.95 per seat.

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• Service Account best practices – Part 2: Least Privilege implementation (0)
• Service Account best practices Part 1: Choosing a Service Account (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)

One typical scenario is that you want to test an application that only shows certain problems if the CPU is under stress. Another typical problem is running a legacy application on a modern computer. Sometimes the old application will run so fast that you are not able to interact with it. Recently, I had such a problem with a new application. I used VMware Workstation on a powerful server, and I wasn’t able to access the BIOS setup of a virtual machine because the “press F2” message was only displayed for a fraction of a second.

Generating CPU load is not as easy as one might think. Of course, you could write a simple script that does the job. For example, a little program that generates random numbers in an endless loop is an option. If you don’t want to write a program yourself, you could use an application like Prime95, a tool for finding Mersenne prime numbers.

Other options are the free tools CPU Burn and CPU Load 2.0. The downside of these tools is that they don’t support multiple processor systems, i.e. they only stress one processor core. This problem can be solved by starting the tool multiple times and assigning the processor affinity via the Windows Task Manager. However, the main disadvantage is that you will only be able to generate a 100% CPU load this way. It might be hard to resume control over your computer, so be careful with these tools. It is probably not a good idea to run them on a productive server.

Cpukiller 3 is a bit smarter. It allows you to configure the processor load you want to generate. It also supports multiple processor cores. You can easily start and stop the resource occupancy via its GUI. The slow down factor can be set with its trackbar. It is also possible to configure hot keys, so you can control Cpukiller easily while you test your application.

Cpukiller 3 costs $29.50, which is a little too expensive in my opinion. I guess this is due to the fact that it has no real competitor. Please, let me know if you know of a cheaper, or even a free, tool that is comparable to Cpukiller. There is a trial which you can use for 30 days. Unfortunately, a trial session is limited to 20 minutes. You have to restart the program to begin a new trial session. So, if you find that Vista’s performance is really too good, then you have no other choice but to buy Cpukiller.

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• FREE: Smart Defrag – A defragmentation tool (9)
• FREE: NovaBench – Free benchmark software (1)

Since quad core CPUs are becoming common, even on desktop computers, there are more options for making use of processor affinity. For example, you could reserve some of your CPU cores for an application that needs more resources. Virtualization solutions are typical candidates for this. Unfortunately, managing processor affinity with Task Manager is a bit cumbersome. That’s where Edgemeal Get and Set CPU Affinity comes in.

The main advantage over Task Manager is that Edgemeal Get and Set CPU Affinity enables you to set the processor affinity for multiple processes. You can easily move all running apps to CPU 0 and your problematic program to CPU 1, for instance. To do this, you just have to mark multiple processes and then assign them to a certain CPU. Of course, there are many other reasons why one might want to change processor affinity.

Another useful feature is the favorites pane. This allows you to save your preferred CPU affinity for certain programs. The next time you want to apply these settings, you can do it with just a mouse click.

If you want your CPU affinity configuration to be set automatically when your computer boots up, then you can set the options “Enable Auto on Start” and “Start with Windows”.

Edgemeal Get and Set CPU Affinity

(Leonardo, thanks for the tip).

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• FREE: Smart Defrag – A defragmentation tool (9)
• FREE: NovaBench – Free benchmark software (1)

Windows Vista is a bloated pig of an operating system. In fact, compared to Windows XP with Service Pack 2 or 3, Vista requires roughly twice the hardware resources to deliver comparable performance.

He knows what he is talking about, because he ran numerous performance tests. I have been discussing some comparable performance tests before, and I don’t want to go into this again, since I think that they are rather pointless.

There is a new Microsoft paper which addresses some of the criticisms that have been put forward against Vista. I find it interesting that MS is quite frank in this text, because usually they only claim that everyone loves Vista. Let me quote from this paper (note: the link is not working at the moment, here (PDF) is a copy):

We‘ve heard some of you say that Windows Vista runs slower than Windows XP on a given PC. So what‘s really happening here? First, we need to avoid comparing apples to oranges—Windows Vista is doing a lot more than Windows XP, and it requires resources to conduct these tasks. That said, it is important to make sure a PC running Windows Vista has enough horsepower to function properly, especially for older PCs running a minimal amount of RAM, since the Windows Vista footprint is larger, bottlenecks will occur.

Well, it is a marketing text, yes. However, in my view, it is a valid argument. I’ve said it before, if Vista were as fast as XP on the same hardware, I would ask for my money back. There is nothing wrong with a “bloated operating system”, just as there is nothing wrong with a “bloated bank account”.

Of course, Kennedy’s point is that Vista has no interesting new features which would justify the additional hardware requirements. A large part of his whole article is about this topic. I am trying to refute this view in my series of blog posts, so I won’t write about this today. However, there is one thing I find remarkable. On the one hand, he claims that Vista requires twice the hardware resources as XP does, and on the other hand, he says that are no real enhancements. It follows that half of the things Vista is doing all day is just a waste of resources.

All in all, there are 10 rounds in his article, but only 4 discuss new Vista features. Obviously, there are many more new features he doesn’t address. Of course, it is impossible to write about all of Vista’s enhancements in just one article. Books have been written about this topic. So I can’t really blame Kennedy for this. However, if you really want to prove that Vista wastes half of the available resources, you have no other choice than to discuss in detail the additional things Vista is doing, and then prove that most of them are useless.

Note that we don’t just talk about Aero or Windows Search. You will certainly regain some of your resources if you disable these. But, even then, Vista will still be slower than XP. So if you have old hardware, which is not powerful enough, you had better let Vista alone. The point is, however, that if we compare both operating systems on a PC that is one or two years old and has at least 2GB RAM, then performance is not an issue at all. Yes, even then, you will be able to measure differences, but they simply don’t matter. It is a common false belief, and even IT pros such as Kennedy are not immune:

For every CPU cycle wasted bringing Vista’s bloated image on par with XP’s, you could be providing your users with an actual performance increase across their core applications.

CPU cycles on desktop computers in a corporate environment usually don’t matter. If your computer uses 30% or 15% of the available CPU resources, it simply makes no difference. There might be a few cases where performance is an issue on an office PC, but for a general argument against Vista you need a little more than that. In the fast majority of cases, you won’t realize the difference. That said, if Vista has just one feature that you find useful and if you have modern hardware, then you can forget everything that has been written about Vista’s performance. Of course, it is another question whether or not you should buy new PCs just because of Vista. This certainly depends on the question of whether your organization benefits from Vista’s enhancements. But this is the topic of the other articles in this series.

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• FREE: Smart Defrag – A defragmentation tool (9)
• FREE: NovaBench – Free benchmark software (1)

I was planning to do some research already for some time, but fortunately the German print magazine C’t (8/2008) just published a couple of comprehensive reports about this topic. So I will summarize some of their findings garnished with my own views. Today, I will focus on performance and in my next posts in this series, I will discuss software compatibility and hardware compatibility.

First of all, why would one prefer Vista x64? I outlined some of the reasons in my earlier post. What I said about security and reliability is still valid. However, a year later things might have changed regarding performance. There are two ways of improving 64-bit operating system’s performance. It utilizes the 64-bit specific computational advantages of the processor and it allows you to use more memory which is often the best way to improve performance.

Computational performance

I suppose that many think that a 64-bit processor performs better than the 32-bit CPU. This might be due to the fact that the transitions from 8-bit to 16 and from 16 to 32 brought some performance gains indeed. At least, this is how I remember it. There are essentially two reasons why a 64-bit Windows should be faster: Compilers have more options for code optimization and all 64-bit CPUs support the so-called SSEx-instructions, a command set that is supposed to be more efficient.

Some 32-bit CPUs support SSEx, too, but most 32-bit apps don’t use them because that would prevent them from working on older 32-bit processors. Hence it follows that only applications which were optimized for 64-bit will show performance gains. Unfortunately, this isn’t the case for most Windows programs, nowadays.

C’t ran a couple of benchmark tests with applications that should most likely benefit from a 64-bit CPU. Such apps are mathematical apps and graphics programs. Their tests show that in most cases Vista x64 isn’t really faster than Vista x86. Sometimes it is even slower. But all in all, the differences are not worth mentioning.

So better CPU performance can hardly be the reason why one would want to move to Vista x64. However, if you have an application that was optimized for 64-bit things look different. Although, Vista has been released more than a year ago, optimized 64-bit apps are still very rare. Note that if your software vendor offers a 64-bit edition, it doesn’t necessarily mean that it will also perform better than the 32-bit app. This just guarantees that it will work under Vista x64.

Memory

Even though CPU performance is not an issue here, performance in general certainly is the main reason that speaks for 64 bit. However, it is not CPU power, but the option to use more memory. In theory, a 32-bit CPU supports 2³² Byte = 4GB RAM. Unfortunately, Vista x86 supports only up to 3.12 GB. For compatibility reasons Vista has to reserve some of its memory for device drivers. So putting more than 3GB in a computer that runs Vista x86 doesn’t make sense.

Also, Vista’s kernel will reserve at least 2GB of the 4GB for its own utilization (including the above mentioned reserve for drivers). That means that you only have 2GB for applications at the most. In most cases that will be enough. Only if you have to run many memory hungry apps simultaneously will you need more.

A disadvantage of 64-bit apps is that they usually need more memory due to 64-bit pointers. Thus, running Vistax64 on a computer with 3GB or less only makes sense if you run applications that were optimized for 64-bit.

The most common reason nowadays why you might need more RAM than 3GB is that you want to run a virtualization solution such as VMware Workstation or Virtual PC on your PC. As long as you run only one guest OS, 3GB might be sufficient. But if you test software where you need a couple of guest systems simultaneously, you certainly have a good reason to move to Vista x64.

Modern laptops with 64-bit CPU support up 4GB and desktops usually allow you to use 8GB. RAM is very cheap now. 4GB sell for less than €100 ($150). So money isn’t a reason anymore for refraining from putting more RAM in your PC. Of course, if you have hundreds or thousands PCs in your organizations that is something different. But then you certainly don’t need a 64-bit OS on all of them. For those PCs that would benefit from Vista x64 the only question is if the notorious software and hardware compatibility problems of Vista x64 are still an issue. I will explore this topic in my next posts.

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• FREE: Smart Defrag – A defragmentation tool (9)
• FREE: NovaBench – Free benchmark software (1)

Note that you have to run this tool with administrator privileges. Under Vista, you must elevate Iometer, that is, it has to be launched explicitly “as administrator”. Iometer’s navigation pane displays the so-called workers. They contain the I/O devices, i.e. the available logical drives. Click on the disk target tab and select one of the drives. You should specify the number of sectors Iometer can use for the performance test. Otherwise it will use all available disk space. For your first test, 1000 sectors will do. Iometer creates a file called iobw.tst in the root folder of the corresponding disk for this purpose.

Next, you have to configure the Access Specifications. You can find them under the corresponding tab. I suggest you use the pre-configured “512B; 100% Read; 0% random” access specification for your first performance test. Add it to the “Assigned Access Specifications” column. Now, click on the “Results Display” tab. Change the update frequency to 3 seconds. This allows you to watch the measurement process more or less in real time.

To start the performance test, you have to click on the green flag in the task bar. Iometer will ask you to specify the name of the csv file where Iomenter will log the results of your performance test. It will then create the iobw.tst file I mentioned above. The status bar will display “Preparing drives” until Iometer is finished. Now you should be able to watch the results of your performance measurement. To display the gauge you see above, you have to click on one of the arrows next to the measured parameters. To end the test, click on the “Stop” icon in the task bar.

Iometer

• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)
• FREE: Smart Defrag – A defragmentation tool (9)
• FREE: NovaBench – Free benchmark software (1)

Most monitoring tools I tried before focus mostly on notification and reporting methods. The problem I have with notifications is that they usually come when I am quite busy with something else. If it isn’t something urgent that requires my immediate attention, I push it aside until I have more time. But I never have more time. If you work as an admin, you most likely know what I am talking about.

The central part of pt360 is its Dashboard. The term dashboard is often misused these days, but this one is a dashboard in its original sense. It gives you a great overview of the states of all essential servers and network devices. And you decide yourself what really is essential. That is, you can configure this dashboard to your own needs. With the new Network Discovery tool, you scan first your whole network environment to get an overview of all your devices. Then you decide which machines need permanent monitoring and add them to your Dashboard.

Basically, there are three kinds of instruments you can attach to your Dashboard: charts, gauges and lists. The “360” in pt360 supposedly is related to the gauges. And I must admit they are my favorite instruments. They don’t only look cool, they are quite useful, too. You can measure a single performance counter with it, for instance the CPU load of a certain server. The charts display readings over a certain time period and the lists show a certain performance counter of multiple machines.

These are the instruments you can attach to your personal Dashboard: Active Directory (via WMI), availability (based on pings), CPU (via SNMP), disk utilization (via SNMP), memory (via SNMP), MS Exchange performance counters (via WMI), network interface performance (via SNMP), web viewer for another browser-based management tool or website (via HTTP), Windows performance counters (via WMI), and MS SQL Server performance counters (via WMI).

The instruments that are based on SNMP (Simple Network Management Protocol) can also monitor non-Windows devices supporting this protocol. WMI (Windows Management Instrumentation) is only for Windows machines, of course. What I like is that PacketTrap’s encrypted credential store consolidates Windows, SSH, FTP and SNMP credentials into a single location. This allows you to monitor devices and computers outside your own Windows domain easily.

If you always have the dashboard open, preferably on a second monitor, you get a feeling of what is normal for your network environment. Hint: Run in full screen mode. The load on your servers depends on many different parameters such as the time of the day, software updates, maintenance, projects, etc. This is why it is so difficult to define thresholds for notifications in advance. But like a good pilot, you know your machinery and can tell with just a glimpse on your dashboard if it’s all hunky-dory.

When you see something suspicious or even alarming, you usually have to run further investigations to find the cause of the problem. That’s where the pt360 Tool Suite comes in. It is a collection of utilities that belongs in every network administrator’s tool box. Although you can get most of these tools for free, it is quite useful to have them all in one user interface. These tools are included in pt360 PRO: Ping Scan, SNMP Scan, DNS Audit, Port Scan, Enhanced Ping, Graphical Ping, WHOIS, WMI Scan, TraceRoute, Traffic Jam, TFTP Server, Wake on LAN, Syslog Server, Cisco config.

The names of the tools speak for themselves. Just let me explain the uncommon ones. Cisco config was added to the Tools Suite recently. It enables you to manage Cisco devices, i.e. you can download, upload, and archive configuration files. You can also recover a lost password with this tool. Enhanced Ping creates reports based on ping results, and Graphical Ping display the response time in a chart.

My favorite tool, however, is Traffic Jam. It is a traffic generator that allows you to stress your network. Like humans, applications behave often differently under pressure. To make sure that your network is also able to cope with heavy loads, you can use Traffic Jam in the evening after work when no productive systems are endangered by your test.

There is not much to criticize about pt360. Even though I prefer continuous network monitoring over notification-based solutions, alerts and reports are certainly important features for every monitoring solution. The version I tested didn’t have these features, but PacketTrap announced that they will support alerts (mail, sms) and reports in April (free upgrade to all pt360 PRO users).

Other new features planned are Time Series Analysis (ability to keep key network data for periods of time as well as run reports on it), Switch Port Mapper (identify which devices are connected to each port on a switch/router), Cisco Netflow (visualizes the flow network traffic), VoIP Analyzer (provides performance statistics including MOS, jitter, network latency, packet loss, and other important QoS metrics).

I tested pt360 PRO on Windows Server 2008 and encountered no problems. The usage is pretty straight forward. I checked the help file once or twice, but most of the features are self-explanatory. pt360 PRO costs $1,499. However, you get a discount of $200 if you buy it in March. There also is a free version, but it lacks many of the high value tools and the Dashboard feature which is certainly an essential part of pt360.

• Poll: Are you currently using a monitoring solution? (11)
• Troubleshoot slow logon – Part 2: The 3-headed monster (0)
• Troubleshoot slow logon – Part 1: Profile size (0)
• SCOM 2012 review – Part 8: Dashboards (0)
• How to disable the Shutdown Event Tracker in Windows Server 2008 R2 (0)