en-us http://www.intensedebate.com/users/295312 Comments by kmwoley http://kmwoley.com/blog/?p=429#IDComment241390730 Yah... wordpress / intensedebate doesn&#039;t do that great of a job with the spam filters sometimes. :) Mon, 19 Dec 2011 04:49:30 +0000 http://kmwoley.com/blog/?p=429#IDComment241390730 http://kmwoley.com/blog/?p=429#IDComment202417543 As I understand it, If Windows won&#039;t boot it without drivers, it is not a full hardware RAID. Fake RAID solutions require drivers. Mon, 3 Oct 2011 05:08:11 +0000 http://kmwoley.com/blog/?p=429#IDComment202417543 http://kmwoley.com/blog/?p=429#IDComment174205411 Yes, you should be able to. It should be pretty easy to test out - pull the drives out and slap them in another machine, they should be recognized just fine. Sat, 16 Jul 2011 16:37:58 +0000 http://kmwoley.com/blog/?p=429#IDComment174205411 http://kmwoley.com/blog/?p=429#IDComment161027790 Good questions - unfortunately, I don&#039;t really know hte answers. I know that mirroring can take quite a while - I forget how long it took my drives to mirror initially, but it was shorter than a day. However, it sounds like you&#039;re striping across 3 disks and I can imagine that reorganizing from 2 to 3 disks could take a very long time. I&#039;m surprised that it was trending towards 2+ weeks, however. I&#039;d be interested in hearing how long it took in the end. Thu, 9 Jun 2011 14:12:29 +0000 http://kmwoley.com/blog/?p=429#IDComment161027790 http://kmwoley.com/blog/?p=429#IDComment161026868 thanks for posting your results! Thu, 9 Jun 2011 14:08:43 +0000 http://kmwoley.com/blog/?p=429#IDComment161026868 http://kmwoley.com/blog/?p=429#IDComment154232010 I have actually heard folks claim that &quot;fake RAID&quot; solutions should perform better - that was one of the reasons that I wanted to run these tests. I wasn&#039;t sure that I believed those claims. And my numbers showed that they weren&#039;t really true - I believe data, not rumor. The word &quot;supposedly&quot; was chosen for exactly that reason. You&#039;re correct in that there is little to no real hardware advantage that fake RAID controllers provide - I&#039;ve heard claims that the drivers that enable the RAID often work better than the RAID provided by the OS, or that &quot;Fake-RAID can have solid performance due to high bandwidth, low latency interfaces to the CPU&quot; due to close ties to the CPU&#039;s south bridge <a href="http:\/\/www.servethehome.com\/difference-hardware-raid-hbas-software-raid\/" target="_blank"> <a href="http:\/\/www.servethehome.com\/difference-hardware-raid-hbas-software-raid\/&lt\;\/a&gt\;" target="_blank"> <a href="http://(http://www.servethehome.com/difference-hardware-raid-hbas-software-raid/)</a>" target="_blank">(http://www.servethehome.com/difference-hardware-raid-hbas-software-raid/)</a></a>. I think it is important to validate these things for one&#039;s self.</a> Fri, 20 May 2011 03:36:19 +0000 http://kmwoley.com/blog/?p=429#IDComment154232010 http://kmwoley.com/blog/?p=429#IDComment154228291 Sure, particularly if you had results that were different from what I found. Fri, 20 May 2011 03:19:07 +0000 http://kmwoley.com/blog/?p=429#IDComment154228291 http://kmwoley.com/blog/?p=560#IDComment128843223 Better viewing angle - there are games that suggest that it be at eye level and I&#039;ve found that it works a little better up there. And it allows for easier worship. Fri, 18 Feb 2011 16:39:45 +0000 http://kmwoley.com/blog/?p=560#IDComment128843223 http://kmwoley.com/blog/?p=429#IDComment124835462 For a RAID 1 config, I&#039;m quite certain you can pull out a disk and slap it into another machine with no problem. Afterall, they&#039;re just mirrored. :) This I did try with my setup. For a RAID 0 config, I&#039;m fairly certain you can transfer the disks from one Windows 7 machine to another, but I have not tried it myself. The key with most RAID configs, as I understand it, is to use the same &#039;controller&#039; that laid down the bits when you move the disks. In a HW RAID 0 config, you&#039;d have to get the exact same controller card to be able to move the disks across machines. For SW raid, it&#039;s the OS doing the controlling so you should need no special hardware - just the same software. I&#039;d love to hear from anyone that&#039;s tried this for themselves. Since I don&#039;t have a RAID 0 config, I can&#039;t prove or disprove it - and I haven&#039;t dug up specific docs on Win 7 saying either way. Mon, 31 Jan 2011 02:20:26 +0000 http://kmwoley.com/blog/?p=429#IDComment124835462 http://kmwoley.com/blog/?p=566#IDComment124832837 Glad I was able to help - thanks for the notes on the changes you made. That&#039;s a good catch. I&#039;ve updated the script to take into account single-lap file directories. FYI - the correct fix is to declare the $datfiles as an array, and then add the new lap files to it. That way the [0] reference will always work later in the script: $datfiles = @(); $datfiles += get-childitem Lap*.dat | Sort-Object; Mon, 31 Jan 2011 01:58:38 +0000 http://kmwoley.com/blog/?p=566#IDComment124832837 http://kmwoley.com/blog/?p=441#IDComment116303515 true - but looking through all the log entries can be a pain. <br />And, the real point of this post was &#39;how to find what you&#39;re looking for&#39; more than the chkdsk logs themselves. :) Fri, 17 Dec 2010 02:16:05 +0000 http://kmwoley.com/blog/?p=441#IDComment116303515 http://kmwoley.com/blog/?p=546#IDComment108281798 We debated on that ordering, too. I&#039;m not sold on which one is right - I just know I say &quot;I don&#039;t hate it&quot; to things I like more than I say &quot;I don&#039;t love it&quot;. Sat, 6 Nov 2010 22:37:20 +0000 http://kmwoley.com/blog/?p=546#IDComment108281798 http://kmwoley.com/blog/?p=429#IDComment108240420 This sort of buffering is important for the sequential read/write testing - it is not as important in the random cases. Note that, in the mixed random tests (the only random test I did), RAID 0 actually performs better. This makes sense because the buffering and spacial locality doesn&#039;t help much, but the independent heads seaking different areas of the two disks should make the random reads faster. Buffering may not be the only reason - but I think it is a good candidate explanation. Sat, 6 Nov 2010 19:19:26 +0000 http://kmwoley.com/blog/?p=429#IDComment108240420 http://kmwoley.com/blog/?p=429#IDComment108240355 I think the difference here is due to the buffering and access pattern. I only did sequential read tests. On RAID 0 (striping), you can get the full advantage of having twice the buffer space on the disks while reading the data back. While Disk 1 is buffering its first 8MB interleaved of data, Disk 2 is also reading/buffer in its 8MB of interleaved data - together, they&#039;ve buffered the first 16MB of sequential data which they can quickly feed to the controller. By contrast, on RAID 1 (mirroring), both disks are reading identical patterns of data. So, while Disk 1 is reading/buffering 8MB of data, Disk 2 is also reading/buffering roughly the same 8MB of data. Thus, those disks together can only buffer about 8MB of sequential data at any moment in time. Sat, 6 Nov 2010 19:18:54 +0000 http://kmwoley.com/blog/?p=429#IDComment108240355 http://kmwoley.com/blog/?p=429#IDComment108236373 Yep - you can try it out yourself. After the RAID has been created on the disks, all of the information for another Win7 machine to read them is contained on the disk themselves. In fact, you can take a single RAID 1 disk and put just the one disk in another machine and it should read just fine (although it will report that it&#039;s partner is missing). My understanding is that this is the case for any RAID implementation - the only difference with HW-based solutions is that you have to get an exact match of your HW controller to read the disk in another machine. Sat, 6 Nov 2010 18:47:20 +0000 http://kmwoley.com/blog/?p=429#IDComment108236373 http://kmwoley.com/blog/?p=429#IDComment97920816 Good points. By &#039;better&#039; I believe I was intending to say &#039;faster&#039;. How much faster, I have no evidence of. I only know what I&#039;ve been told, since I&#039;ve never owned a hardware RAID solution myself. All I do know is that the very few people I know who have spent a few hundred $$ on hardware RAID cards sware by them with respect to performance. In theory, it does make sense, particularly if you&#039;re running CPU bound workloads and you want to offload the (little) processing that a software RAID solution incurs to the card. I completely agree with you that software RAID has big advantages in terms of not being tied to specific hardware. That&#039;s one of it&#039;s biggest selling points for me. Fri, 10 Sep 2010 01:22:40 +0000 http://kmwoley.com/blog/?p=429#IDComment97920816 http://kmwoley.com/blog/?p=429#IDComment97156884 I believe you can only Mirror (RAID 1) the partition that Windows is on if you&#039;re using Windows to do the RAIDing. Sun, 5 Sep 2010 16:33:07 +0000 http://kmwoley.com/blog/?p=429#IDComment97156884 http://kmwoley.com/blog/?p=429#IDComment97156624 It&#039;s most likely a software/hardware RAID and Windows likely installed the drivers for it automatically. Real raid controllers typically cost 10x the fake raid versions, so you generally know what you&#039;re paying extra for. Sun, 5 Sep 2010 16:31:20 +0000 http://kmwoley.com/blog/?p=429#IDComment97156624 http://kmwoley.com/blog/?p=429#IDComment93480286 Thanks for the feedback. Glad it was useful for you. I&#039;d be interested in seeing someone else run these performance numbers and report back to see if they have similar results. Kevin Mon, 16 Aug 2010 17:19:25 +0000 http://kmwoley.com/blog/?p=429#IDComment93480286 http://kmwoley.com/blog/?p=429#IDComment93479506 Thanks for the feedback. Glad it was useful for you. I&#39;d be interested in seeing someone else run these performance numbers and report back to see if they have similar results. <br />Kevin Mon, 16 Aug 2010 17:13:32 +0000 http://kmwoley.com/blog/?p=429#IDComment93479506