Alright, I really don't like asking for help, but I'm desperate. I'm the Dir of IT for a mid-size company with around 8 Windows servers and less than 1000 computers running Windows 7. We're mostly power users around here however, I've yet to figure out a solution to deal with fragmentation - primarily on the servers. I've been a Windows user since the early days and it's hard to believe that these issues still exist with the Windows file system. Listen, I would take a more stable file system over the flashy graphics in Windows 8 any day. Bottom line - our employees are workaholics and work remotely at all hours. I can't run a manual or even schedule a defrag job on these servers without interfering significantly with the performance of all the applications we run. Take Exchange - Microsoft recommends running the defragmentation process during off-hours and preferable with the Exchange databases stopped (http://support.microsoft.com/kb/328804). Really? We don't have "off-hours" - is this really my only option!?! Take our file servers - they get pretty hammered with the constant data transfers going on - with the high rate of file editing we do around here. What does Microsoft recommend here? Well, the defragmentation utility on Windows 2008 is TURNED OFF by default - so what does that tell you? I should run it manually right? Not if I want to keep my job... If I sound frustrated - it's because I am. What does the rest of the IT community do? Peter

Hi, It looks like you need a high availability solution to keep your workflow going at a constant pace and to service the users of the system The Cheap Way: Its simple, schedule downtime well in advance and notify all involved to work around it The Other Way: Sadly this is not so cheap. It would involve replicating your environment on to backup servers and storage in which you would fail over to when maintenance needs to take place. This would involve an identical server setup and a datastore running in synchronous mode. When you need to perform maintenance, hold IO for 2-3 mins and fail over to the backup system. Your company cannot have it both ways, they need to consider that downtime is often unavoidable, all you can do as a IT person is find the best way to minimise that downtime window. Martin If you find my information useful, please rate it. :-)

Hi, It looks like you need a high availability solution to keep your workflow going at a constant pace and to service the users of the system The Cheap Way: Its simple, schedule downtime well in advance and notify all involved to work around it The Other Way: Sadly this is not so cheap. It would involve replicating your environment on to backup servers and storage in which you would fail over to when maintenance needs to take place. This would involve an identical server setup and a datastore running in synchronous mode. When you need to perform maintenance, hold IO for 2-3 mins and fail over to the backup system. Your company cannot have it both ways, they need to consider that downtime is often unavoidable, all you can do as a IT person is find the best way to minimise that downtime window. Martin If you find my information useful, please rate it. :-)

Your environment is exactly why we created Diskeeper. As you point out, the effects of file and free space fragmentation are far reaching. Every file stored and referenced by Windows can be affected. When an application needs to read/write data and the file system has determined that it spans more than a single range of LCNs, then a "split I/O" happens for each extent of the file. This increases the amount of I/O traffic sent to the storage system. Each I/O takes a measurable amount of time and this is one of the major causes of poor application performance and behavior. This is true for both physical and virtual systems. It is also not addressed or handled by SAN storage as some manufactures have falsely stated. Waiting around and allowing fragmentation to build up and then address it in the traditional way isn't the ideal solution. I am sure you would agree that a better apporach is to be proactive and prevent fragmentation from occuring in the first place. This avoids the overhead of the traditional break/fix defrag aspect and gives you the performance improvement right at the time a file is being created/modified. When a file does need to be moved, Diskeeper does it without taking system resources away from other improtant activities. There is no need for you to stop Exchange or force users to log out to see the benefits Diskeeper provides. In fact, Diskeeper (now Condusiv Technologies) is listed as a Microsoft Exchange Partner. Here's some information you will find useful... http://www.condusiv.com/disk-defrag/email-servers/ http://downloads.condusiv.com/pdf/Exchange_Servers_Run_Like_New.pdf http://www.condusiv.com/products/diskeeper/ Thanks, Howard

I'm sorry Martin, but I dont think your suggestions will work for me: Scheduling downtime you mean get fired? This idea doesnt exist in my workplace. If its something critical thats a one-time deal, then fine. But trying to explain the need for all of our systems to go download regularly for a couple of hours to defrag seems a little much, no? Replicating my environment with all due respect, if I would put money into something, it wouldnt be this. I refuse to believe that the only way to deal with fragmentation is to duplicate my entire infrastructure! Why dont we get two offices, and swap between them as the need arises? I understand the idea of failover, but for handling fragmentation ridiculous! Again, with all due respect, money is not an issue here its lack of a realistic solution that will making its way across the top brass approval line over here. As for Howard, thank you so much for sharing this information. Defragging Exchange without unmounting the Message Store - that's pretty friggin' unbelievable! Got Diskeeper Server trialware running through its paces on a Dev box over here will keep you posted! Peter