Maticks Posted August 18, 2018 Share Posted August 18, 2018 So been stable for about a month since replacing my LSI Card and all the drive read errors went away. I've since told all my drives to stay spinning and not shutdown. Today one drive threw a few read errors. i ran smart on it and it seems fine. It only did it once for 64 read errors. i am not sure if the drive is dying or is it my Corsair RM850 Power supply that is about 5 years old but i only us about 200Watts out of the 850. It is also a 5 year old hard drive so it could be end of life for this drive, it is also not a NAS drive but a Desktop drive before they had NAS drives at a decent price. Smart shows two errors below. Not sure if i should replace the hard drive or power supply here. And if i did replace the power supply whats the best for lots of SATA Hard Drives, i only have a 4770 i7 processor it doesn't eat a lot of power. ATA Error Count: 2 CR = Command Register [HEX] FR = Features Register [HEX] SC = Sector Count Register [HEX] SN = Sector Number Register [HEX] CL = Cylinder Low Register [HEX] CH = Cylinder High Register [HEX] DH = Device/Head Register [HEX] DC = Device Command Register [HEX] ER = Error register [HEX] ST = Status register [HEX] Powered_Up_Time is measured from power on, and printed as DDd+hh:mm:SS.sss where DD=days, hh=hours, mm=minutes, SS=sec, and sss=millisec. It "wraps" after 49.710 days. Error 2 occurred at disk power-on lifetime: 44771 hours (1865 days + 11 hours) When the command that caused the error occurred, the device was active or idle. After command completion occurred, registers were: ER ST SC SN CL CH DH -- -- -- -- -- -- -- 40 51 00 ff ff ff 0f Error: UNC at LBA = 0x0fffffff = 268435455 Commands leading to the command that caused the error were: CR FR SC SN CL CH DH DC Powered_Up_Time Command/Feature_Name -- -- -- -- -- -- -- -- ---------------- -------------------- 25 00 00 ff ff ff ef 00 10d+01:05:33.157 READ DMA EXT 25 00 00 ff ff ff ef 00 10d+01:05:33.156 READ DMA EXT 25 00 00 ff ff ff ef 00 10d+01:05:33.156 READ DMA EXT 25 00 00 ff ff ff ef 00 10d+01:05:33.155 READ DMA EXT 25 00 00 ff ff ff ef 00 10d+01:05:33.155 READ DMA EXT Error 1 occurred at disk power-on lifetime: 44771 hours (1865 days + 11 hours) When the command that caused the error occurred, the device was active or idle. After command completion occurred, registers were: ER ST SC SN CL CH DH -- -- -- -- -- -- -- 40 51 00 ff ff ff 0f Error: UNC at LBA = 0x0fffffff = 268435455 Commands leading to the command that caused the error were: CR FR SC SN CL CH DH DC Powered_Up_Time Command/Feature_Name -- -- -- -- -- -- -- -- ---------------- -------------------- 25 00 00 ff ff ff ef 00 10d+01:05:28.874 READ DMA EXT 25 00 00 ff ff ff ef 00 10d+01:05:28.874 READ DMA EXT 25 00 00 ff ff ff ef 00 10d+01:05:28.873 READ DMA EXT 25 00 00 ff ff ff ef 00 10d+01:05:28.593 READ DMA EXT 25 00 80 ff ff ff ef 00 10d+01:05:28.583 READ DMA EXT Num Test_Description Status Remaining LifeTime(hours) LBA_of_first_error # 1 Short offline Completed without error 00% 44785 - # 2 Short offline Completed without error 00% 43113 - # 3 Extended offline Completed without error 00% 39712 - # 4 Extended offline Completed without error 00% 34033 - # 5 Extended offline Completed without error 00% 33394 - # 6 Extended offline Completed without error 00% 9088 - # 7 Extended offline Interrupted (host reset) 50% 7329 - # 8 Short offline Completed without error 00% 7154 - # 9 Extended offline Completed without error 00% 6588 - #10 Short offline Completed without error 00% 6409 - #11 Extended offline Completed without error 00% 5918 - #12 Short offline Completed without error 00% 5738 - #13 Extended offline Completed without error 00% 5174 - #14 Short offline Completed without error 00% 4994 - #15 Extended offline Completed without error 00% 4431 - #16 Short offline Completed without error 00% 4250 - #17 Extended offline Completed without error 00% 3713 - #18 Short offline Completed without error 00% 3530 - #19 Extended offline Completed without error 00% 2965 - #20 Short offline Completed without error 00% 2786 - #21 Extended offline Completed without error 00% 2288 - Link to comment
pwm Posted August 18, 2018 Share Posted August 18, 2018 It's impossible to say - the only thing the HDD says is that it can't read this block. Trying twice will just make it fail twice. But what we can't know is if this is because of a physical problem with the drive, or if there was vibrations or similar when this block was last written. By the way - SMART short tests will not scan the surface, so they will normally not fail unless there is a huge error with the drive. Link to comment
Maticks Posted August 18, 2018 Author Share Posted August 18, 2018 just started an extended test does that do a surface check? not 100% sure here if its just an old drive. considering just changing out the power supply it is 5 years old i can put it in my garage as a spare. But these old Desktop Seagate drives don't have any vibration sensors or resistance to vibration at all. not really designed for NAS. Link to comment
Maticks Posted August 18, 2018 Author Share Posted August 18, 2018 i do have File Integrity installed i could hit the drive with a disk check that would at least verify if a file fails checksum. Not to mention put the drive under high read. Link to comment
pwm Posted August 18, 2018 Share Posted August 18, 2018 1 hour ago, Maticks said: just started an extended test does that do a surface check? Yes. 1 hour ago, Maticks said: considering just changing out the power supply There is nothing indicating any error with the PSU. 1 hour ago, Maticks said: i do have File Integrity installed i could hit the drive with a disk check that would at least verify if a file fails checksum. Definitely a good idea to see if you have broken files - but don't start this test until after having run an extended test. The two tests would interfere with each other. Link to comment
JorgeB Posted August 18, 2018 Share Posted August 18, 2018 I'm guessing this is a Seagate disk, since this type of error is common with them, I'd say if it passes the extended test (and it should) give it a second chance, but any more similar errors I would replace it. Link to comment
Vr2Io Posted August 18, 2018 Share Posted August 18, 2018 Read your previous post, RM850 was a good PSU. Your have 13 drive, how many cable-set plug to PSU ( fully module ? ) for those disk, if one or two, it should not good enough. Total peripheral/SATA socket at PSU side was 5 and SATA cable-sets provide should be 3. Pls seperate them in no more 4 disk per socket and in good connection status. ** Suggest you unplug those cable in PSU side and check connector pin have burn-out or not ** Link to comment
Maticks Posted August 18, 2018 Author Share Posted August 18, 2018 21 minutes ago, johnnie.black said: I'm guessing this is a Seagate disk, since this type of error is common with them, I'd say if it passes the extended test (and it should) give it a second chance, but any more similar errors I would replace it. Yeah its a seagate disk, but its a desktop drive and 5 years operational time. so i have to be expecting a death at some point soon. WD Red drives i have had mixed results with, Seagate NV NAS drives seem to be ok. i dont know which one is better. but lets see what this extended smart test shows up then ill run the file integrity check see if i have some corrupted files or not. After my multi disk issue with the LSI Card i decided to roll out file integrity so i could see what was corrupted if anything. Link to comment
JorgeB Posted August 18, 2018 Share Posted August 18, 2018 I wasn't implieing Seagates are bad, just that specific error UNC at 0x0ffffff, which is a bogus LBA address, only seems to happen to them. Link to comment
Maticks Posted August 18, 2018 Author Share Posted August 18, 2018 11 minutes ago, Benson said: Read your previous post, RM850 was a good PSU. Your have 13 drive, how many cable-set plug to PSU ( fully module ? ) for those disk, if one or two, it should not good enough. Total peripheral/SATA socket at PSU side was 5 and SATA cable-sets provide should be 3. Pls seperate them in no more 4 disk per socket and in good connection status. ** Suggest you unplug those cable in PSU side and check connector pin have burn-out or not ** Thats the PSU I have. I have 6 Disks on one cable-set, 4 Disks on another cable set, and a Molex cable connector in the Peripheral section with 2 Hard Drives on a Molex to Sata with 3 Case Fans on the Molex as well. I have 2xslots for SATA power leads so another 8 Disks directly from the PSU. These PSU's are single rail at least, i did some research thinking i had a multi rail issue, i have an RM850i as well for my gaming system and spare SATA power connectors for that seems to look the same i was going to do a test with an old 1TB disk to see if it powers up. it should be the same pin out. I moved to these things to clean up my cabling, but i actually think they have something in them that is causing my problem, they note to have 2200μF capacitors to help stable power. They look great nice and clean but i am suspect of them. The ebay 2xSATA Y Spliters ive not had issues with. https://www.silverstonetek.com/product.php?pid=304 Link to comment
Maticks Posted August 18, 2018 Author Share Posted August 18, 2018 5 minutes ago, johnnie.black said: I wasn't implieing Seagates are bad, just that specific error UNC at 0x0ffffff, which is a bogus LBA address, only seems to happen to them. yeah these drives always make Clicky noises as well when they park. its really unnerving when you have 3-4 of them. I've rolled out a few over the years in other systems its normal. WD's seem to just die in my experience, Seagate seem to operate and throw some errors here and there before hand. but i have had one disk throw some errors like this and still working 2 years later without any issues. Link to comment
Vr2Io Posted August 18, 2018 Share Posted August 18, 2018 22 minutes ago, Maticks said: Thats the PSU I have. I have 6 Disks on one cable-set, 4 Disks on another cable set, and a Molex cable connector in the Peripheral section with 2 Hard Drives on a Molex to Sata with 3 Case Fans on the Molex as well. I have 2xslots for SATA power leads so another 8 Disks directly from the PSU. These PSU's are single rail at least, i did some research thinking i had a multi rail issue, i have an RM850i as well for my gaming system and spare SATA power connectors for that seems to look the same i was going to do a test with an old 1TB disk to see if it powers up. it should be the same pin out. I moved to these things to clean up my cabling, but i actually think they have something in them that is causing my problem, they note to have 2200μF capacitors to help stable power. They look great nice and clean but i am suspect of them. The ebay 2xSATA Y Spliters ive not had issues with. https://www.silverstonetek.com/product.php?pid=304 You means have 20 disk which power from this PSU ? Not understand this -> "I have 2xslots for SATA power leads so another 8 Disks directly from the PSU." How could you directly connect as the PSU was full module. 6 disks on one cable-set really bad. Take another cable-set from RM850i and test it by a old disk was perfect. Capacitors inside a cable won't got problem but also won't help much. Link to comment
Maticks Posted August 18, 2018 Author Share Posted August 18, 2018 14 minutes ago, Benson said: You means have 20 disk which power from this PSU ? Not understand this -> "I have 2xslots for SATA power leads so another 8 Disks directly from the PSU." How could you directly connect as the PSU was full module. 6 disks on one cable-set really bad. Take another cable-set from RM850i and test it by a old disk was perfect. Capacitors inside a cable won't got problem but also won't help much. I will be bring the system to 20 disks in the future i have 2 more module slots for cables into the power supply for future disks. Link to comment
pwm Posted August 18, 2018 Share Posted August 18, 2018 59 minutes ago, Maticks said: I have 6 Disks on one cable-set Assuming 2 Ampere start current, that means you may reach 12 A start current for that cable. You should avoid going higher than 8-10 A for a single power pin. Link to comment
Vr2Io Posted August 18, 2018 Share Posted August 18, 2018 49 minutes ago, Maticks said: I will be bring the system to 20 disks in the future i have 2 more module slots for cables into the power supply for future disks. Take care about 5v current rating for 20 disks, RM850's 3.3v+5v combine should be 150w, minus 50w of system used, 100w for disks ( i.e. 5v, 20A) was just enough even whole system use far more less then 850w. Link to comment
Maticks Posted August 18, 2018 Author Share Posted August 18, 2018 Since i am putting more disks on a power supply is their one better suited for SATA powered drives on 5Volts? Most of these consumer PSU's seem to be angled at PCI E power connections. Link to comment
John_M Posted August 18, 2018 Share Posted August 18, 2018 21 minutes ago, Maticks said: Most of these consumer PSU's seem to be angled at PCI E power connections. Many consumers have more graphics cards than hard disks these days. I'd buy two more SATA power cables and split the load across the five outlets - four disks on each would equal your required 20. Fans don't draw much current so I'd use the motherboard headers to power them. The big current draw is on the +12 volt rail as disks spin up. I don't believe 3.5-inch hard disks draw much current on the +5 volt rail but I'm struggling to find a data sheet to confirm this. Both WD Red and Seagate Ironwolf spec sheets neglect to include the +5 volt current rating but give a start up current of around 1.75 to 2.0 amps (depending on the model) on the +12 volt rail. Link to comment
Frank1940 Posted August 18, 2018 Share Posted August 18, 2018 10 minutes ago, John_M said: I don't believe 3.5-inch hard disks draw much current on the +5 volt rail but I'm struggling to find a data sheet to confirm this. I looked at the label on a Seagate Ironwolf 3TB and the 5V current rating is .55A. I find it hard to believe that the current draw for any modern for the electronics is much different. Link to comment
Vr2Io Posted August 18, 2018 Share Posted August 18, 2018 6 disks won't got problem on 5v usage, but 20 disks will be another story. Nowadays PSU just for gaming need, not count storage need. Someone say those combine power was enough, the rating on disk spec. just the max. Anyway you may do some math. Link to comment
Vr2Io Posted August 18, 2018 Share Posted August 18, 2018 3 hours ago, Maticks said: Since i am putting more disks on a power supply is their one better suited for SATA powered drives on 5Volts? Most of these consumer PSU's seem to be angled at PCI E power connections. Even 1000w+ PSU, the 3.3+5v also 100~150w only. Link to comment
Frank1940 Posted August 18, 2018 Share Posted August 18, 2018 Actually, you have to do the math for all of the voltages to make sure you don't exceed any of the maximum power ratings. I have see a lot of 450W PS's where the maximum current rating on a single rail 12V buss was in the 36A range. A quick bit of Math will show that if you draw that much, you can't have any load on any of the other outputs! Remember if you exceed any rating for more than a millisecond, the PS will go into its shutdown mode. This isn't to protect the wiring, it is to protect the components in the regulating circuits. Link to comment
John_M Posted August 18, 2018 Share Posted August 18, 2018 Sorry, but I don't understand the problem with the +5 volt rail. Doing the mathematics and using @Frank1940's figure of 0.55 amp for each disk drawn from the +5 volt rail: Total current drawn from +5 volt rail by 20 disks = 20 x 0.55 amps = 11 amps Total power drawn from +5 volt rail by 20 disks = 11 amps x 5 volts = 55 watts 13 hours ago, Benson said: RM850's 3.3v+5v combine should be 150w, minus 50w of system used, 100w for disks ( i.e. 5v, 20A) So what's the problem? The +5 volt rail has a lot of capacity (around 45 watts) to spare. Link to comment
pwm Posted August 19, 2018 Share Posted August 19, 2018 1 hour ago, Frank1940 said: Remember if you exceed any rating for more than a millisecond, the PS will go into its shutdown mode. Many better PSU allows a significant short-term overload - the switch transistors etc can handle quite a lot of current to handle load spikes but it's mostly the cooling that specifies the maximum allowed load. And since the PSU power components and fitted heatsinks have thermal mass, it takes a while for the temperature to rise high enough that the overload circuits will step in. It's also quite common for both single-rail and multi-rail units that the current overload cut-off is 30-50% over the specified current. One reason for allowing short-term overloads is to handle the large current spike on power-on, when all the empty capacitors has to be charged. Some of this inrush current is handled by slow-start features where the PSU ramps up the voltages. But even with slow-start, the initial current can be quite significant. The following review tested the RM850 at 110% load with proper function - so the PSU obviously does not instantly shuts down. It does have over-power protection, but it seems +10% isn't enough to trig that protection. If overloading a single rail, they could most probably have run at way more than 10% overcurrent without shutdown - at least for a limited time until the thermal protection steps in. https://www.techpowerup.com/reviews/Corsair/RM850/5.html In the end, it's normally the thermal protection that will make a PSU shutdown on overload - the short-circuit protection normally steps in at a very large overcurrent. Link to comment
Vr2Io Posted August 19, 2018 Share Posted August 19, 2018 7 hours ago, John_M said: Sorry, but I don't understand the problem with the +5 volt rail. Doing the mathematics and using @Frank1940's figure of 0.55 amp for each disk drawn from the +5 volt rail: Total current drawn from +5 volt rail by 20 disks = 20 x 0.55 amps = 11 amps Total power drawn from +5 volt rail by 20 disks = 11 amps x 5 volts = 55 watts So what's the problem? The +5 volt rail has a lot of capacity (around 45 watts) to spare. 22 hours ago, Benson said: Take care about 5v current rating for 20 disks, RM850's 3.3v+5v combine should be 150w, minus 50w of system used, 100w for disks ( i.e. 5v, 20A) was just enough even whole system use far more less then 850w. The words in begin and end not say must have problem. Below are one ** selected ** power hunger disk model ( HGST 7K6000 ) http://www.hgst.com/sites/default/files/resources/Ultrastar_7K6000_SATA_512n_OEM_Specification_Rev1.0.pdf You will found Random peak max @0.85 and Sequential peak @0.77, if by math it still not max at PSU. But would you design / planning in this way ? This is your choose. Pls also note, all PSU when run in its max rating, different kind of degrade also need consider. http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story4&reid=369 i.e. when 5v draw 18A, then voltage will drop to 4.95v. Link to comment
John_M Posted August 19, 2018 Share Posted August 19, 2018 I haven't assumed anything. I've just used the figures I was given because I was struggling to find them in the dumbed down spec sheets. Re-doing the calculation, using the highest value quoted (0.85 amp): Total current drawn from +5 volt rail by 20 disks = 20 x 0.85 amps = 17 amps Total power drawn from +5 volt rail by 20 disks = 17 amps x 5 volts = 85 watts which is still comfortably less than the 100 watts you say is available for disks. So, again, I don't see a problem. Are you saying that to run at 85% of available capacity is bad? If so, what do you recommend? Remember that this is peak current, not average. 6 hours ago, Benson said: when 5v draw 18A, then voltage will drop to 4.95v 5.00 volts - 4.95 volts = 0.05 volt = 50 mV, which represents a 1% drop. How is that even significant when it is so much smaller than the tolerance on the rail voltage anyway? Link to comment
Recommended Posts
Archived
This topic is now archived and is closed to further replies.