Rajahal

Mine got into the 40s using the stock fanplate with 9 drives. Ambient was probably around 26.

That's a Norco 4220, right? Did you mod the fanplate yourself? Looks like 9x80mm fans - what are you drive temps like? And how loud is it?

Right. Just don't ever turn it on.

Why? There's no need. The stock 80 mm fans work fine for the back. I'm considering replacing them just to make the server as quiet as possible, but that's the only reason. Theoretically a good quality 450W PSU like the Corsair 450VX is enough to power 20 green drives. I'm not sure if that has actually been tested in practice, though. When powering a massive amount of drives, the amps on the +12V rail are more important than the wattage. The Antec 520W I'm using has 40A on the +12V rail. Each green drive should use around 2A. 40/2 = 20 drives, so I should be fine as long as I stick with all green drives. So I'll have to phase out my 7200 rpm drives as I get closer to the 20 drive limit. Or swap out the PSU for a bigger one. Your PSU may or may not be overkill depending on if you use 7200rpm drives, or if your PSU has much lower amps on the +12V rail.

Finally finished modding my Norco 4220 fanplate. I cut out the excess steel with an angle grinder, then used a hand file and a bench grinder to smooth the edges. I then drilled new screw holes with a drill press. It wasn't too much trouble since I had the right tools for the job, but the $15 pre-made plate from Norco is probably a better option (once it becomes available). One of the 120mm fans (the one of the left in the 2nd picture) is attached by only two screws. The other side hangs off the edge of the fan plate. It is still plenty secure, I don't hear any vibration or other problems. The power cables that it rubs up against help keep it stable as well. I also chose to move the fan control board into the motherboard area for easier access (its attached with Liquid Nails). The 120 mm fans do a much better job of keeping the drives cool. They all stay in the 30s now - 5400rpm drives will be between 30C and 35C, whereas the 7200rpm drives will be between 35C and 39C, with the exception of my single 7200rpm 1.5 TB Seagate, which sometimes creeps up to 41C. I think that drive just likes to run a bit hot, since I've seen it in the low 40s in other cases as well. The server is leaps and bounds quieter. It still isn't silent, you can definitely hear it running. However, its no longer loud enough to interfere with media enjoyment. Parts list: Case: NORCO RPC-4220 Mobo: ASUS P5BV-M CPU: Intel Celeron 430 RAM: 2 GB Kingston DDR2 800 (low profile) PSU: Antec NEO ECO 520C 520W Continuous Power ATX12V v2.3 / EPS12V 80 PLUS Certified Active PFC Power Supply Cables: 4 x NORCO C-SFF8087-D and 1 x 3ware CBL-SFF8087OCR-06M Fans: 3 x COOLER MASTER R4-S2S-124K-GP 120mm Case Fan Add-on Card: SuperMicro AOC-SASLP-MV8 HDDs: Parity: 2TB Seagate LP Data: 2TB WD EARS, 1.5TB WD EARS, 1.5TB Seagate 7200rpm, 1TB Samsung F1 7200rpm, 1TB WD EARS, 2 x 1TB WD EADS Cache: 320 GB Seagate 7200rpm Total storage: 9 TB Total capacity: 38 TB Now I just need to add a second SuperMicro AOC-SASLP-MV8 and this server will be complete with a 20 drive capacity. Since I don't have the need for extra drive bays at the moment (I have 12 bays hooked up and only 9 are used), I figure I will wait and try to catch the add-on card on sale. Assuming I can find the second SuperMicro card for about $100, my total cost for this build (without hard drives) is $739.52. I'm not sure it is currently possible to beat that price, since I got many of these parts on sale. The motherboard was a free gift with the case, so that's a savings of at least $100. The RAM I got for $22 back in the day. The SuperMicro card was $97. The PSU was $30 after a rebate. The CPU, fans, and cables were just about the only things I paid full price for, thought the 3ware cable was an open box deal.

Yes, it is absolutely necessary for unRAID. Best case scenario, you are seeing decreased performance. Worst case scenario, you will encounter corrupted data. No, the jumpers don't come with the drives but they are cheap and you can scavenge them from old drives. See the wiki for more info. ~~~~~~~~~~~~~~~ Attached are the preclear results for a 5900 RPM 2 TB Seagate (haven't upgraded the firmware yet), and two 1.5 TB WD EARS drives with jumpers. I saw some worrying stuff on all of them, but I've misinterpreted this stuff before, so I would appreciate a second opinion. 2_TB_Seagate_2x1.5_TB_WD_EARS_Preclear_results_8-22-2010.txt

Great, thanks for the good news Joe. Edit: I determined that it is the Promixe TX4 card that has both of it's drives in IDE mode. The card has no BIOS screen of its own, and I can't figure out how to control it via the motherboard's BIOS settings (all of which are correct, by the way). I plan on migrating my server to my 4220 + AOC-SASLP-MV8 setup soon, so I guess I won't worry about it.

Well Joe, I need your advice. Attached is the syslog after I precleared a brand new 2 TB WD EARS w/ jumper. It took 64 hours to complete! I figure something must be wrong for it to take that long. First off, I know I have HPA on some of my drives. I no longer use a gigabyte mobo and I haven't bothered to get rid of them. Some things I've noticed: Jul 28 18:36:38 Tower emhttp: pci-0000:00:11.0-scsi-0:0:0:0 host1 (sda) WDC_WD10EADS-00L5B1_WD-WCAU45951367 Jul 28 18:36:38 Tower emhttp: pci-0000:00:11.0-scsi-1:0:0:0 host2 (sdb) WDC_WD10EADS-00L5B1_WD-WCAU49026373 Jul 28 18:36:38 Tower emhttp: pci-0000:00:11.0-scsi-2:0:0:0 host3 (sdc) WDC_WD10EARS-00Y5B1_WD-WMAV50673569 Jul 28 18:36:38 Tower emhttp: pci-0000:00:11.0-scsi-3:0:0:0 host4 (sdd) SAMSUNG_HD103UJ_S13PJ1LS631384 Jul 28 18:36:38 Tower emhttp: pci-0000:00:14.1-ide-0:0 ide0 (hda) WDC_WD15EARS-00Z5B1_WD-WMAVU2570536 Jul 28 18:36:38 Tower emhttp: pci-0000:00:14.1-ide-0:1 ide0 (hdb) WDC_WD20EARS-00MVWB0_WD-WMAZA0132941 Jul 28 18:36:38 Tower emhttp: pci-0000:03:05.0-scsi-0:0:0:0 host5 (sde) ST3320620AS_6QF0WRZW Jul 28 18:36:38 Tower emhttp: pci-0000:03:05.0-scsi-1:0:0:0 host6 (sdf) WDC_WD5000AAJS-22TKA0_WD-WCAPW5380433 Two of my drives (one of which is this new 2 TB WD EARS) are in IDE emulation mode even though I have AHCI turned on in the BIOS. Any idea why? I'm using a Promise TX4 card, do I need to do some special configuration for that? Jul 30 10:22:15 Tower kernel: hdb: hdb1 Jul 30 10:22:25 Tower kernel: udev: starting version 141 Jul 30 10:46:06 Tower kernel: hdb: dma_intr: status=0x51 { DriveReady SeekComplete Error } Jul 30 10:46:06 Tower kernel: hdb: dma_intr: error=0x40 { UncorrectableError }, LBAsect=98144375, sector=98144112 Jul 30 10:46:06 Tower kernel: hdb: possibly failed opcode: 0x25 Jul 30 10:46:06 Tower kernel: end_request: I/O error, dev hdb, sector 98144112 Jul 30 10:46:06 Tower kernel: Buffer I/O error on device hdb, logical block 12268014 Jul 30 10:46:06 Tower kernel: Buffer I/O error on device hdb, logical block 12268015 Jul 30 10:46:06 Tower kernel: Buffer I/O error on device hdb, logical block 12268016 Jul 30 10:46:06 Tower kernel: Buffer I/O error on device hdb, logical block 12268017 Jul 30 10:46:06 Tower kernel: Buffer I/O error on device hdb, logical block 12268018 Jul 30 10:46:06 Tower kernel: Buffer I/O error on device hdb, logical block 12268019 Jul 30 10:46:06 Tower kernel: Buffer I/O error on device hdb, logical block 12268020 Jul 30 10:46:06 Tower kernel: Buffer I/O error on device hdb, logical block 12268021 Jul 30 10:46:06 Tower kernel: Buffer I/O error on device hdb, logical block 12268022 Jul 30 10:46:06 Tower kernel: Buffer I/O error on device hdb, logical block 12268023 Jul 30 10:46:34 Tower kernel: hdb: dma_intr: status=0x51 { DriveReady SeekComplete Error } Jul 30 10:46:34 Tower kernel: hdb: dma_intr: error=0x40 { UncorrectableError }, LBAsect=98144375, sector=98144112 Jul 30 10:46:34 Tower kernel: hdb: possibly failed opcode: 0x25 Jul 30 10:46:34 Tower kernel: end_request: I/O error, dev hdb, sector 98144112 Jul 30 10:46:34 Tower kernel: __ratelimit: 54 callbacks suppressed Jul 30 10:46:34 Tower kernel: Buffer I/O error on device hdb, logical block 12268014 Jul 30 10:46:34 Tower kernel: Buffer I/O error on device hdb, logical block 12268015 Jul 30 10:46:34 Tower kernel: Buffer I/O error on device hdb, logical block 12268016 Jul 30 10:46:34 Tower kernel: Buffer I/O error on device hdb, logical block 12268017 Jul 30 10:46:34 Tower kernel: Buffer I/O error on device hdb, logical block 12268018 Jul 30 10:46:34 Tower kernel: Buffer I/O error on device hdb, logical block 12268019 Jul 30 10:46:34 Tower kernel: Buffer I/O error on device hdb, logical block 12268020 Jul 30 10:46:34 Tower kernel: Buffer I/O error on device hdb, logical block 12268021 Jul 30 10:46:34 Tower kernel: Buffer I/O error on device hdb, logical block 12268022 Jul 30 10:46:34 Tower kernel: Buffer I/O error on device hdb, logical block 12268023 Jul 30 20:03:59 Tower kernel: hdb: dma_intr: status=0x51 { DriveReady SeekComplete Error } Jul 30 20:03:59 Tower kernel: hdb: dma_intr: error=0x40 { UncorrectableError }, LBAsect=1907719879, sector=1907719496 Jul 30 20:03:59 Tower kernel: hdb: possibly failed opcode: 0x25 Jul 30 20:03:59 Tower kernel: end_request: I/O error, dev hdb, sector 1907719496 Jul 30 20:03:59 Tower kernel: __ratelimit: 54 callbacks suppressed Jul 30 20:03:59 Tower kernel: Buffer I/O error on device hdb, logical block 238464937 Jul 30 20:03:59 Tower kernel: Buffer I/O error on device hdb, logical block 238464938 Jul 30 20:03:59 Tower kernel: Buffer I/O error on device hdb, logical block 238464939 Jul 30 20:03:59 Tower kernel: Buffer I/O error on device hdb, logical block 238464940 Jul 30 20:03:59 Tower kernel: Buffer I/O error on device hdb, logical block 238464941 Jul 30 20:03:59 Tower kernel: Buffer I/O error on device hdb, logical block 238464942 Jul 30 20:03:59 Tower kernel: Buffer I/O error on device hdb, logical block 238464943 Jul 30 20:03:59 Tower kernel: Buffer I/O error on device hdb, logical block 238464944 Jul 30 20:03:59 Tower kernel: Buffer I/O error on device hdb, logical block 238464945 Jul 30 20:03:59 Tower kernel: Buffer I/O error on device hdb, logical block 238464946 ...a bunch of these Buffer I/O errors on hdb... Jul 31 10:43:07 Tower preclear_disk-diff[12912]: S.M.A.R.T. error count differences detected after pre-clear Jul 31 10:43:07 Tower preclear_disk-diff[12912]: note, some 'raw' values may change, but not be an indication of a problem Jul 31 10:43:07 Tower preclear_disk-diff[12912]: 54c54 Jul 31 10:43:07 Tower preclear_disk-diff[12912]: < 1 Raw_Read_Error_Rate 0x002f 100 253 051 Pre-fail Always - 0 Jul 31 10:43:07 Tower preclear_disk-diff[12912]: --- Jul 31 10:43:07 Tower preclear_disk-diff[12912]: > 1 Raw_Read_Error_Rate 0x002f 200 200 051 Pre-fail Always - 2 Jul 31 10:43:07 Tower preclear_disk-diff[12912]: 65c65 Jul 31 10:43:07 Tower preclear_disk-diff[12912]: < 197 Current_Pending_Sector 0x0032 200 200 000 Old_age Always - 0 Jul 31 10:43:07 Tower preclear_disk-diff[12912]: --- Jul 31 10:43:07 Tower preclear_disk-diff[12912]: > 197 Current_Pending_Sector 0x0032 200 200 000 Old_age Always - 2 Jul 31 10:43:07 Tower preclear_disk-diff[12912]: 67c67 Jul 31 10:43:07 Tower preclear_disk-diff[12912]: < 199 UDMA_CRC_Error_Count 0x0032 200 253 000 Old_age Always - 0 Jul 31 10:43:07 Tower preclear_disk-diff[12912]: --- Jul 31 10:43:07 Tower preclear_disk-diff[12912]: > 199 UDMA_CRC_Error_Count 0x0032 200 200 000 Old_age Always - 0 ...and then two read errors and two pending sectors. Looks like I may need to RMA this drive? Syslog-7-31-2010-2TBEARSpreclear..txt

Your cable management is simply beautiful. Well done.

This looks like a match: YMI Aluminum 5-in-3 SATA II /SAS HDD Enclosure

I see your point, barry. In terms of general computer design in which the CPU and graphics card are the hottest components, then your advice is spot on. However, look at kizer's server - there are pictures of it on the previous page of this thread. I believe the stock fan that comes with the Sempron 140 is more than adequate to keep it cool with no other help from the system fans. There are also vents immediately above and behind the CPU. Add to that the case fan located behind the CPU, and I think you'll agree that it has no chance of overheating. The onboard video card is also underutilized, and doesn't get hot. Therefore, the only concern (in my mind) is the hard drives. In kizer's current configuration, all the hard drives are in the lower bays of the case with no dedicated fan. Therefore, the closest fan is the one found in the PSU. The case fan near the CPU most likely isn't powerful enough to pull air all the way from the opposite corner of the case (especially since there are so many other vents and places for air to be pulled). This is why I was advising kizer to flip the PSU around so that it's fan could help pull air across the hard drives. Air would then enter from the lower drive bays across the hard drives and be sucked out the back by the PSU fan. Once kizer fills all the hard drive bays, he may need to add a fan or two to the case's top vents to help cool the drives (unless he uses hot swap cages or drive cages with dedicated fans). I'm having to recall kizer's setup from memory since my work's firewalls block the image host he used. I hope I'm remembering it correctly. Now that I think about it, the CM 590's internal drive cage does come with a dedicated fan, so maybe he's using that. I wish I could see those pictures (care to rehost them on imgur or imageshack, kizer?).

Why would you want that to happen? You want hot air to exit your case as soon as possible. Your temps are perfectly fine, though, so again no need to change anything at the moment.

Looks great Kizer! I'm glad to see so many people benefiting from the Budget Box design. At first I was surprised that you were able to route the 4 pin mobo power cable behind the mobo tray, but then I realized that you installed the PSU with the fan facing down. It is generally recommended that you install the PSU with the fan facing up so that it can suck hot air out of the case and blow it out the back. The downside of this is that the 4 pin mobo power cable will then be too short to route behind the mobo tray. See the pic in my post up further on this page for an example. However, as long as you don't have any problems with heat in your server, then it probably isn't worth the trouble to flip the PSU around.

For those wondering, here's the parts list for the 15 bay server I built for guiri. CPU: AMD Sempron 140 Mobo: Biostar A760G M2+ RAM: 1 GB Kingston DDR2 800 Case: Cooler Master Centurion 590 PSU: Corsair CMPSU-400CX Hot Swap 5-in-3's: 5 in 3 BP-SATA350 (red) (x 3) Flash Drive: SanDisk Cruzer Micro 2GB Flash Drive (USB2.0 Portable) Model SDCZ6-2048-A11 unRAID license: Pro SATA expansion card 1: SuperMicro AOC-SASLP-MV8 SATA expansion card 2: 2 Port SATA Serial ATA PCI RAID Controller Card - Silicon Image (only one port used) SAS Cables: 3ware Serial Attached SCSI (SAS) Internal Cable - 1 x SFF-8087 - 4 x SATA - 1.6ft (x 2) Misc: SATA cables, HDD jumpers Notice the similarity to the recommended budget build I also shipped the server with 6 pre-installed and pre-cleared hard drives. I used the cache disk slot to reserve one of the drives as a warm spare (but disabled the cache disk function on all user shares). The server was completely configured, all guiri had to do was plug in the flash drive and power and network cables. I also provide very thorough documentation, getting started instructions, pictures, animations, etc with every server I sell. As for the temperature of that one enterprise drive (47 C), I agree that it is a bit too hot. However, I believe that once guiri replaces it with a low power green drive, that the problem will be solved. I have advised him to use only green drives in this server, as the 400 W PSU and heating limitations of the server design won't support hot, fast drives like the 2 TB enterprise drive he is currently using when the server is maxed out at 15 drives. If simply replacing the drive doesn't solve the problem, then I agree with the others that an extra fan or two in the top slots should take care of it. I am back from Europe. If anyone else is interested in a custom server, PM or email me. I would especially like to start making some small 3-6 drive servers designed for the unRAID Plus license so that I'm not directly competing with LimeTech (since LimeTech also makes and sells 12-15 drive servers) (see here for an example). These servers will of course be much cheaper than this 15 bay server.

The server has to be booted into unRAID to run preclear, so I know unRAID was up, however I don't remember if the array was started or not. I'm guessing not, since it wouldn't have had any disks assigned anyway, since I was preclearing all of them (this was a brand new server I was building for guiri, another member here). The CPU was the Sempron 140, which is a single core 2.7 GHz. 900 MHz is pretty slow, maybe that's the reason. Also, no need to 'hack' the script. Joe L. included many commands that you can append to the initial preclear command that will change what the script does: -n = Do NOT perform preread and postread of entire disk to allow SMART firmware to reallocate bad blocks in the clearing process -N = Do not perform read validation during postread. (skip this step) (basic test to check if values read are all zero as expected. Skipping this test will save a few miniutes, but possibly not detect a drive that returns non-zero values when zeros were expected as bad.) -c count = perform count preread/clear/postread cycles where count is a number from 1 through 20 If not specified, default is 1 cycle. For large disks, 1 cycle can take 10 or more hours -t = Test if disk has pre-clear signature. This option may NOT be combined with the -c or -n options. The test does not write to the disk. It makes no changes to a disk at all. It only reads the first 512 bytes of the disk to verify a pre-clear signature exists. Note: "-t" does not read the entire disk to verify it it pre-cleared as that could take hours for a large disk. since the pre-clear-signature is written *after* a disk is entirely filled with zeros, if it exists, we assume the disk is cleared. -w size = write block size in bytes -r size = read block size in bytes -b count = number of blocks to read at a time -v = print version of ./preclear_disk.sh -m [email protected] = optional recipient address. If blank and -M option is used, it will default to default e-mail address of "root" -M 1 = Will send an e-mail message at the end of the final results (default if -m is used, but no other -M option given) -M 2 = Will send an e-mail same as 1 plus at the end of a cycle (if multiple cycles are specified) -M 3 = Will send an e-mail same as 2 plus at the start and end of the pre-read, zeroing, post-read -M 4 = Will send an e-mail same as 3 plus also at intervals of 25% during the long tests The -m, -M options requires that a valid working mail command is installed. One version that has worked (bashmail) is affilaited with the unraid_notify script. There are others that also will work. Any of these need to be configured to work with your mail server. The unraid_notify script will have instructions on how to configure this. See http://lime-technology.com/forum/index.php?topic=2470.0 for unraid_notify and http://lime-technology.com/forum/index.php?topic=2961.0 for the mail script. NOTE: The lastest verion of mail has to be used. earlier versions of the mail script affiliated with unraid_notify do not support the standard mail syntax needed. Unless the -n option is specified the disk will first have its entire set of blocks read, then, the entire disk will be cleared by writing zeros to it. Once that is done the disk will be partitioned with a special signature that the unRAID software will recognize when the drive is added to the array. This special signature will allow the unraid software to recognize the disk has been pre-cleared and to skip an initial "clearing" step while the server remains off-line. The pre-read and post-read phases try their best to exercise the disk in a way to identify a drive prone to early failure. It performs reads of random blocks of data interspersed with reads of sequential blocks on the disk in turn. This program also uses non-buffered reads of the first and last cylinders on the disk, the goal is to perform those reads in between the others, and to keep the disk head moving much more than if it just read each linear block in turn. Take from here. Though reading through the above there doesn't seem to be a command that runs just the post read. I'm sure Joe could help you with that if needed.

Yeah, that does seem slow. My latest preclear results were from 2 TB WD EARS Greens (w/ jumpers), and they each took 35 hours. I precleared 4 simultaneously (using a mixture of onboard and PCIe slots), and the speed didn't change. Server had only 1 GB of RAM.

Nice! That is one monstrous case...looks like you could get close to 20 drives if you went for all 5-in-3's. You would probably need a bigger PSU at that point, though.

gfjardim, would you mind using a different host for your pictures? The one you are using doesn't get past my work's firewall (and I expect I'm not the only one who browses these forums from work ). I recommend imgur.

Wow, that sure is compact. Nice cabling too. No heat issues?

...and here's the preclear results of the last two disks. Everything looks normal to me, but again, I would appreciate a second set of eyes confirming this. guiri_syslog_5-30-2010_-_precleared_1.5_TB_EADS_and_last_2_TB_EARS.txt

Just want some confirmation that the attached preclear results for 4 different drives are OK (3x 2TB WD EARS w/ jumpers, 1x 1.5TB Seagate 7200 rpm). Everything looked normal to me, but I just wanted to make sure I didn't miss anything. I had to view this syslog in Wordpad for some reason. The formatting is all kinds of messed up when I open it with Notepad. Weird. guiri_syslog_-_5-25-2010_-_preclear_3_x_2_TB_+_1_x_1.5_TB.zip

So thats 9 x 80 mm fans? Wow! Great work! What type of fans are you using? I assume you won't use the ones that came with the 4220. Once you get it all set up, I would appreciate you posting the drive temperatures you see. I also have a 4220 and I would like to make it quieter eventually.

Nice rig, unraided! Just to clarify, are those disk enclosures trayless? It sounds like they are not. Love your cable management! That must have taken some work.

yumbrad: What motherboard are you running? It seems to have a LOT of SATA ports. Also, what's with the IDE drive taped on top? Is that a cache drive or something? Being taped on top may lead to heat issues...

Great tip, thank you! That may be just what I need to get my drive temps down to acceptable levels this summer. I would much rather sacrifice those panels than the dust filters themselves.