WEBVTT 00:00:07.120 --> 00:00:14.160 So, I just had a bit of a scare, guys. I thought Murphy's Law really had it in 00:00:11.599 --> 00:00:18.800 for me today. This is the stack of eight 3 TBTE hard drives that I'm using to 00:00:16.240 --> 00:00:23.600 upgrade my storage server that I archive all my footage on and all of that. And 00:00:21.279 --> 00:00:28.000 uh it appeared to me not that long ago, you can see the network is critical, 00:00:25.760 --> 00:00:32.000 which is bad. It appeared to me that I had lost two hard drives at the same 00:00:29.840 --> 00:00:37.520 time, meaning that the redundancy that Windows Home Server V1 provides um was 00:00:35.280 --> 00:00:42.320 not applicable to any data that was on both of the failed hard drives. Um 00:00:40.320 --> 00:00:46.879 however, it looks like upon further inspection, the Western Digital 1 TB 00:00:44.640 --> 00:00:53.120 Black that's in there is back up and running and only the Hitachi right here 00:00:49.680 --> 00:00:55.120 is actually dead. So, um, that was a 00:00:53.120 --> 00:00:59.760 real relief because it means that I do have to, uh, repair the network or the, 00:00:58.399 --> 00:01:06.080 uh, the backup database in all likelihood. Oh, no. It looks like the backup database is okay. Awesome. Those 00:01:03.280 --> 00:01:09.799 were on the WD. Um, but I do have to, uh, I do have to remove the Hitachi 00:01:08.240 --> 00:01:16.240 drive at some point. So, yeah, that's 00:01:13.720 --> 00:01:20.000 um, very disappointing and very frustrating because it takes a few hours 00:01:18.000 --> 00:01:24.560 to get that done. And I was really hoping to get these new drives in there 00:01:22.400 --> 00:01:29.119 and get them uh get them rated up and get Windows Home Server V2 on here. I 00:01:26.560 --> 00:01:33.280 had to This is just packaging for the new drives. Had to pull out some 00:01:31.600 --> 00:01:37.040 existing drives. This one failed a little while ago. The Wildfire was just 00:01:35.119 --> 00:01:44.560 in there for testing purposes. There's a few Seagates. These are old Seagates. 00:01:40.840 --> 00:01:47.280 7200.10s. All of them survived. Um which 00:01:44.560 --> 00:01:53.520 is a testament to these particular drives, I guess. two 320s and a 250. So, 00:01:49.840 --> 00:01:56.320 those are kicking it. Well, kicking 00:01:53.520 --> 00:02:00.719 back, not working anymore. Well, they work, but they're not going to continue 00:01:58.000 --> 00:02:04.399 to work for me. Um, whatever. You guys get the point. So, this Hitachi is going 00:02:02.399 --> 00:02:07.920 to come out. Uh, that WD is going to move slots, and I'm going to be putting 00:02:06.240 --> 00:02:11.039 Oops, sorry. I'm going to be putting the eight new drives in the eight bays at 00:02:09.840 --> 00:02:15.120 the bottom. And then I'm going to be going with uh, you know what? Maybe I'll 00:02:12.959 --> 00:02:19.200 throw the Wildfire in. It's got toggle nan, so it should be pretty reliable for 00:02:17.280 --> 00:02:23.840 the uh boot drive of Windows home server v2 or 2011, whatever you guys want to 00:02:21.680 --> 00:02:27.680 call it. Veil. Um, it's been brought to my attention that you don't have to use 00:02:25.440 --> 00:02:31.200 a 240 gig drive. You can get away with a 120 with a little edit during the 00:02:29.760 --> 00:02:34.879 installation process. So, that's a really good thing. And, uh, oh yeah, 00:02:33.360 --> 00:02:38.879 right. The kicker for all of this was that when the two drives were out, it 00:02:36.879 --> 00:02:46.400 told me the backup database was failed. And um I actually just bricked the OS of 00:02:43.760 --> 00:02:52.000 my wife's computer and was about to use the home server backup restore utility 00:02:49.760 --> 00:02:55.040 to get her computer back up and running. So I thought I had lost pretty much 00:02:53.440 --> 00:02:59.200 everything. But now that that one drive is working, I'm in uh pretty good shape. 00:02:57.360 --> 00:03:02.560 So thanks for coming along for the ride, guys. And I'll keep you posted on my 00:03:00.400 --> 00:03:05.920 Windows home server upgrade. Don't forget to subscribe to Linus Tech Tips 00:03:04.080 --> 00:03:11.440 for more unboxings, reviews, and other excuse me, other computer videos. 00:03:09.440 --> 00:03:16.560 So, none of the new drives got detected at all. They're all detecting as zero 00:03:13.680 --> 00:03:21.200 gigabytes, which um stands to reason since the firmware I'm running on my 00:03:18.080 --> 00:03:23.760 controller is older than my cats. So, 00:03:21.200 --> 00:03:28.440 I'm uh updating the firmware. All I got to do apparently is 00:03:25.879 --> 00:03:30.760 this. 00:03:30.760 --> 00:03:37.280 And apparently that didn't work. So, I'll 00:03:35.920 --> 00:03:43.200 give it another crack. I'll get it I'll get it updated and then we'll see how things go once we get booted into back 00:03:40.799 --> 00:03:50.040 into Windows and create the array. I think I'm going to go with a RAID 00:03:45.959 --> 00:03:52.319 6. That worked. The file name just got 00:03:50.040 --> 00:03:57.760 truncated. So, this is my first boot after updating the firmware. This is 00:03:54.799 --> 00:04:03.040 new. I hope that's a good sign. All right. So, I'm into my RAID 00:04:00.200 --> 00:04:08.480 configuration. Physical drives. Let's see if they Oh, they are detected now. 00:04:06.799 --> 00:04:12.280 All right. So, I guess we might as well do a quick tutorial on how to create a 00:04:11.200 --> 00:04:18.400 RAID volume on an Ara RAID card. So, we're 00:04:16.000 --> 00:04:21.320 going to call this uh 00:04:21.320 --> 00:04:25.919 RAID six. I don't have another RAID six. 00:04:24.720 --> 00:04:32.040 All my other drives are just pass through drives, which just means they're 00:04:28.000 --> 00:04:36.040 standalone drives. RAID set was created 00:04:32.040 --> 00:04:36.040 successfully. Cool. 00:04:38.040 --> 00:04:44.759 Um, so let me 00:04:41.240 --> 00:04:48.960 see. Okay, so I could expand 00:04:44.759 --> 00:04:50.800 it. I could hm activate incomplete raid 00:04:48.960 --> 00:04:56.560 set. I guess that's pretty much it. I could create hot spares. I can rescue 00:04:53.040 --> 00:04:56.560 raid sets. Delete hot 00:04:57.000 --> 00:05:02.800 spares. Oh, neat. That's actually not a bad idea. I should probably use one as a 00:05:01.360 --> 00:05:06.639 hot spare since I don't really need all the capacity to go uh to go with it 00:05:05.280 --> 00:05:12.440 right now. So, what a hot spare will do is if a drive fails, it'll automatically 00:05:09.280 --> 00:05:15.240 go right in and rebuild the 00:05:12.440 --> 00:05:21.280 uh rebuild the array. 00:05:17.560 --> 00:05:21.280 So, let's have a 00:05:21.880 --> 00:05:29.680 look at the actually Oh, no, not this one. Sorry. 00:05:26.400 --> 00:05:29.680 Let's have a look 00:05:29.880 --> 00:05:36.960 at the volume that has just been created. Disc management. Here we 00:05:41.400 --> 00:05:45.199 go. Uh, 00:05:46.840 --> 00:05:50.800 refresh. Rescan discs 00:05:50.919 --> 00:05:59.400 maybe. There it is. No, wait. That's not it. 20 gigs. Oh, let's see if we can 00:05:56.960 --> 00:06:03.840 find it. I wonder if the OS is even compatible. Haven't done this in a 00:06:01.440 --> 00:06:10.680 while. So, uh, for one thing, I screwed up when I created it, and I accidentally 00:06:06.080 --> 00:06:14.000 created it with only, uh, seven drives. 00:06:10.680 --> 00:06:15.919 So, okay, there we go. Now, it has 00:06:14.000 --> 00:06:20.000 member discs, eight out of eight. Now, we have to create a volume set. So, we 00:06:18.160 --> 00:06:26.800 create select the RAID set to create a volume set. Then we make a volume name 00:06:22.880 --> 00:06:29.840 and we're going to call it RAID six 00:06:26.800 --> 00:06:31.680 again. Okay. Volume RAID level. This is 00:06:29.840 --> 00:06:36.600 where we can actually edit the uh change the RAID level. Volume capacity maximum 00:06:35.280 --> 00:06:40.720 18 terab. Excellent. 00:06:41.319 --> 00:06:48.560 Um yeah, these are 00:06:44.919 --> 00:06:50.160 4K. Foreground initialization should be 00:06:48.560 --> 00:06:54.170 fine. Let's go with default for all this 00:06:52.840 --> 00:06:56.199 stuff. 00:06:56.199 --> 00:07:02.840 Okay. Volumes to be created 00:06:59.400 --> 00:07:06.639 one. Here we go. Volume set has been 00:07:02.840 --> 00:07:09.199 created. Now we should be able to see it 00:07:06.639 --> 00:07:09.199 in disk 00:07:10.919 --> 00:07:16.620 management in theory. 00:07:18.560 --> 00:07:23.520 theories don't always work out that way. Give me a bit. Ah, yes, it's 00:07:21.759 --> 00:07:27.759 initializing. I'll be back once it's done. That takes a while. All right, 00:07:25.039 --> 00:07:32.440 there we go. It is in a RAID state normal now, which means that I can go 00:07:30.319 --> 00:07:38.880 ahead and disc management. Aha, welcome to the 00:07:36.240 --> 00:07:45.199 initialize and convert disc wizard. Next, disk 16 00:07:42.199 --> 00:07:46.120 initializing. Finish. 00:07:45.199 --> 00:07:52.479 So there is my 00:07:48.599 --> 00:07:54.160 16 terbte volume which has been split up 00:07:52.479 --> 00:07:58.639 and I forget how this works. Yeah, we need to convert to a GPT disc so that we 00:07:56.800 --> 00:08:04.400 can make it the full size instead of being limited. 00:08:01.560 --> 00:08:10.520 So can be only be accessed from Windows server blah blah blah. Okay, got it. 00:08:08.960 --> 00:08:17.440 Primary partition assign X for 00:08:14.440 --> 00:08:17.440 extreme. 00:08:17.560 --> 00:08:26.000 Next grade. Whoops. Apparently I have caps 00:08:21.919 --> 00:08:27.960 lock on already. Raid six. Perform a 00:08:26.000 --> 00:08:33.039 quick format. 00:08:30.440 --> 00:08:34.580 Finish. Because the cluster count is higher than 00:08:35.959 --> 00:08:40.000 expected. That's interesting. 00:08:43.880 --> 00:08:50.480 Yeah. Why don't we try not a quick Oh, that's going to take forever. Let's try 00:08:48.000 --> 00:08:54.920 one more time. Oh, okay. Well, let's see if I can figure this out now. Got to 00:08:52.320 --> 00:09:00.560 love extreme hardware. Always just works. Found a great uh article on the 00:08:58.560 --> 00:09:06.160 support site for Microsoft for the default cluster sizes for NTFS. And it 00:09:03.279 --> 00:09:11.279 looks like even though my volume is greater than 16 terabytes, it is not 00:09:09.200 --> 00:09:14.240 defaulting to 8 kilobytes. So, as soon as I recreate 00:09:17.240 --> 00:09:24.959 it using an 8 00:09:20.839 --> 00:09:27.800 kilobyte, here we go. Setting. We should 00:09:24.959 --> 00:09:35.040 be able to get access to the drive formatting and healthy local disc 00:09:32.000 --> 00:09:36.480 X. There we go. So, now let's run a 00:09:35.040 --> 00:09:42.399 quick benchmark and find out how fast this RAID 6 is. Now, we've all been 00:09:38.720 --> 00:09:45.360 spoiled by SSDs when it comes to huge 00:09:42.399 --> 00:09:48.680 ADO scores, but uh I'm still pretty optimistic so far. Looking at this guy 00:09:47.839 --> 00:09:54.000 right here, holy 00:09:51.080 --> 00:10:01.680 cow, we've already reached 1 Gigabyte per second in sustained reads at 00:09:57.560 --> 00:10:04.240 16K. At 30K, we're up to over 1.2 00:10:01.680 --> 00:10:08.399 gigabytes per second reads. The rights are slower because we're going to be 00:10:05.680 --> 00:10:13.040 controller limited on those. So, as fast as your RAID controller is is as fast as 00:10:10.320 --> 00:10:18.360 you can write to a RAID six. Um, a RAID five would be faster on the rights. Holy 00:10:15.680 --> 00:10:22.680 smokes. We're up over 1.6 gigs per second. And it looks like that's 00:10:20.640 --> 00:10:27.920 probably where we're going to peak. So, 00:10:24.680 --> 00:10:29.920 wow. 1.5 on that one. So, so yeah, we 00:10:27.920 --> 00:10:36.320 peak at around 700 megs per second, right? and around 1.5 or one pix gigs 00:10:34.000 --> 00:10:39.519 per second read. Just ridiculous. Okay, I'll be back once that the benchmarks. 00:10:38.000 --> 00:10:44.800 So, there you go, guys. That's what we ended up with. Uh, now let's do another 00:10:41.760 --> 00:10:46.160 run at a deeper Q depth. So, that should 00:10:44.800 --> 00:10:51.839 give us some interesting results. These are just staggering staggering numbers 00:10:48.720 --> 00:10:54.079 for a mechanical setup. Well, not much 00:10:51.839 --> 00:10:58.160 of an impact on scores. I am curious though to see how this array performs in 00:10:56.399 --> 00:11:03.279 RAID five as opposed to RAID six. So, I'll try RAID 5 with a hot spare, which 00:11:00.560 --> 00:11:07.040 basically gives similar data protection to RAID six because you could have two 00:11:05.040 --> 00:11:10.640 drives fail as long as they don't fail at exactly the same time and the hot 00:11:08.720 --> 00:11:14.959 spare would swoop right in and take over for the one that failed, whereas RAID 6 00:11:12.320 --> 00:11:18.079 can take two failures at the same time. I just want to see how much of a 00:11:16.560 --> 00:11:21.519 performance difference we see in these right performance in these right 00:11:19.920 --> 00:11:26.399 performance numbers with RAID 5. So, thank you for checking out this little 00:11:23.040 --> 00:11:29.040 RAID 6 experiment and uh stay tuned for 00:11:26.399 --> 00:11:32.800 more on my Windows home server upgrade. Don't forget to subscribe to Linus Tech 00:11:30.720 --> 00:11:32.800 Tips.