July 16, 201114 yr Hi guys, Long time lurker, first time poster.. I've recently added 8 new drives to my unRAID box and since then I can't seem to run the box for more than a few hours before the parity drive goes red. I've even bought a NEW parity drive and have the same issue. I am running uTorrent on a network machine that's doing i/o to the box. Nothing too crazy. I have a 650 WATT PSU. 13 drives are WD green drives and the other 7 drives are a mix of WD BLACK/BLUE and 1 maxtor drive. Could the failure be related to lack of power? I never had this issue before I added the 8 additional drives. Those 8 drives are running off of a Supermicro AOC-SASLP-MV8 Marvell 6480 8 Channel SAS/SATA RAID PCI-E Card which was sitting idle in my machine before then. Temps on the box are low to mid 40's The syslog is here: http://dl.dropbox.com/u/11559988/syslog.txt. Really not sure how to start troubleshooting this issue. Any help is greatly appreciated.
July 16, 201114 yr Hi guys, Long time lurker, first time poster.. I've recently added 8 new drives to my unRAID box and since then I can't seem to run the box for more than a few hours before the parity drive goes red. I've even bought a NEW parity drive and have the same issue. I am running uTorrent on a network machine that's doing i/o to the box. Nothing too crazy. I have a 650 WATT PSU. 13 drives are WD green drives and the other 7 drives are a mix of WD BLACK/BLUE and 1 maxtor drive. Could the failure be related to lack of power? I never had this issue before I added the 8 additional drives. Those 8 drives are running off of a Supermicro AOC-SASLP-MV8 Marvell 6480 8 Channel SAS/SATA RAID PCI-E Card which was sitting idle in my machine before then. Temps on the box are low to mid 40's The syslog is here: http://dl.dropbox.com/u/11559988/syslog.txt. Really not sure how to start troubleshooting this issue. Any help is greatly appreciated. what specific make/model power supply? 13 * 2 = 26 Amps + 7 * 3 = 21 Amps 26 + 21 + a few more amps for fans, motherboard, etc... you need a single rail 12V supply of a least 50 Amps.
July 16, 201114 yr Check the cabling to the parity drive. Reseat SATA and power connectors. What type of PSU?
July 16, 201114 yr Author Thanks for the replies so far. The PSU is an Antec TP Trio 650: http://www.antec.com/specs/TP3_650_spe.html I'm guessing it has 3 rails which may be what's causing the problem. If it is in fact the PSU, would you guys have any good suggestions for a PSU that would be ideal for this setup? Thanks!
July 16, 201114 yr Thanks for the replies so far. The PSU is an Antec TP Trio 650: http://www.antec.com/specs/TP3_650_spe.html I'm guessing it has 3 rails which may be what's causing the problem. If it is in fact the PSU, would you guys have any good suggestions for a PSU that would be ideal for this setup? Thanks! Your supply has 4 independent 12 volt rails. They are each capable of 19 Amps, as shown in the page you supplied the link to. They are allocated as follows as described in the user manual: http://www.antec.com/pdf/manuals/TP-650_Manual_EN.pdf 12 Volt Rail 1 (20+4) motherboard main connector 1, Molex Connectors from PSU, SATA Power connectors from PSU 12 Volt Rail 2 4-pin ATX12V (these are the connectors used to supply the power to the CPU on the motherboard) 8-pin EPS12V (these are the connectors used to supply the power to the CPU on the motherboard) 12 Volt Rail 3 PCI-E w/ blue stripe 12 Volt Rail 4 PCI-E w/ green stripe (modular) So.. of the 4 rails, only 1 of them is used to power all the disks, and fans, AND it uses that same rail to supply the 24 pin connector on the motherboard. Rail 2 is used to power your CPU. Rail 3 and 4 are probably un-used, and not usable as they typically do not have the required 5 volts to run a disk drive. They are specifically designed for power hungry PCIe video cards. you are lucky the server is spinning up the drives at all. The one 12 volt rail is rated for 19 Amps. On spin up, you are probably drawing near 50 amps from it. I'd say you have the equivalent of a 300 Watt supply, because of the 4-rail design with 2 unused rails. There is a whole thread on power supply selection. http://lime-technology.com/forum/index.php?topic=12219.0 You want one with a SINGLE high-current 12 volt rail. In your case, with as many disks as you have, you need at least 50 amp capacity. Joe L.
July 16, 201114 yr Author Hello Joe, thank you for the reply. I just did some digging around on that thread and it looks like I need to get a new PSU ASAP! However, would this be the reason that my parity drive keeps failing? I am going to go ahead and change the PSU, but I'm just wondering if you think this has contributed to the constant parity failures. Thanks!
July 17, 201114 yr Hello Joe, thank you for the reply. I just did some digging around on that thread and it looks like I need to get a new PSU ASAP! However, would this be the reason that my parity drive keeps failing?Yes I am going to go ahead and change the PSU, but I'm just wondering if you think this has contributed to the constant parity failures. It is the most likely cause. In fact, the drives you replaced might have been perfectly fine if they were not damaged by the low voltage because the existing supply is rated at about one fourth the capacity it needs.
July 17, 201114 yr According to this review at Jonnyguru this is a single rail power supply disguised as a multi-rail power supply: http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story2&reid=1 This does not mean that the power supply is not the problem. I would suspect cabling, one of your new hard disks or your SASLP though.
July 19, 201114 yr Author Hi guys, Just an update: I purchased a new power supply which, I would say, is more than adequate for the hardware I'm running. It's a SeaSonic X Series X-850 (SS-850KM Active PFC F3) found below: http://www.newegg.com/Product/Product.aspx?Item=N82E16817151102 I did a parity check which took 24+12 hours and came home today from work to find the parity drive red again. I opened up the syslog and found a lot of write errors as well as many other miscellaneous ones that I need some help figuring out. Please find attached a new syslog for your review, if any of you can help out. Any help figuring this out once and for all is greatly appreciated. Thanks again. http://dl.dropbox.com/u/11559988/syslog.txt
July 20, 201114 yr It appears that your system was unable to write to the parity drive and it diabled it. This is what causes the red dot. Jul 19 13:43:17 Tower kernel: ata9.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen Jul 19 13:43:17 Tower kernel: ata9.00: failed command: WRITE DMA EXT Jul 19 13:43:17 Tower kernel: ata9.00: cmd 35/00:48:ef:e1:96/00:00:e0:00:00/e0 tag 0 dma 36864 out Jul 19 13:43:17 Tower kernel: res 40/00:ff:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout) Jul 19 13:43:17 Tower kernel: ata9.00: status: { DRDY } Jul 19 13:43:17 Tower kernel: ata9: hard resetting link Jul 19 13:43:27 Tower kernel: ata9: softreset failed (device not ready) Jul 19 13:43:27 Tower kernel: ata9: hard resetting link Jul 19 13:43:37 Tower kernel: ata9: softreset failed (device not ready) Jul 19 13:43:37 Tower kernel: ata9: hard resetting link Jul 19 13:43:48 Tower kernel: ata9: link is slow to respond, please be patient (ready=0) Jul 19 13:44:12 Tower kernel: ata9: softreset failed (device not ready) Jul 19 13:44:12 Tower kernel: ata9: limiting SATA link speed to 1.5 Gbps Jul 19 13:44:12 Tower kernel: ata9: hard resetting link Jul 19 13:44:17 Tower kernel: ata9: softreset failed (device not ready) Jul 19 13:44:17 Tower kernel: ata9: reset failed, giving up Jul 19 13:44:17 Tower kernel: ata9.00: disabled I do not think the fact your parity drive is being disabled is significant. Try swapping the SATA cable with another drive (I think your parity is attached to the motherboard? Swap with a drive on a controller card.) You will have to revisit the devices page and assign the drives to the correct slots. Make sure that the cable and controller are both swapped. If you can get the problem to move, then that will help narrow things down. (I understand that this problem has happened with different drives in the parity slot, so I think it is safe to eliminate the parity drive for now.) It's always a good idea to run a memtest. Have you run one overnight recently? I would be surprised if the problem was memory, but it is pretty easy to check.
July 20, 201114 yr Author ohlwiler- Thank you for the suggestions! Before I got a chance to read your thread I did the following: Swapped the motherboard, ram, processor, onboard vid and USB flash with my roommates. He was previously running unRAID with no issues with the same configuration, we're in the process of combining our storage space so he had the stuff laying around. As of now the only variables that have NOT been changed on this box is the Norco back plane for the drives and the sata cables. The box is currently doing another parity rebuild. If I get failures after this I will try swapping the sata cable, and if that fails I guess the only constant would be the Norco backplane. Thanks again for the support thus far guys!
Archived
This topic is now archived and is closed to further replies.