Re: dodgy hd
On Tue 16 Jun 2009 22:50:18 NZST +1200, Steve Holdoway wrote: > Until there is a standard and requirement, you can't *guarantee* it will > work or rely on the results. It'll *probably* work, there are plenty of > cases where it plain doesn't work... like my SSD here smart doesn't necessarily make a lot of sense for SSDs, and they weren't mentioned so far. The questions here were "is my disk stuffed" and "it's stuffed, do I claim warranty". Both plentifully answered. > Until that time, smartmon is a really useful tool, but it's not the sole > solution you make it out to be. Well then you're even more out of luck, because step 1 still is: check your hardware first. Volker -- Volker Kuhlmann is list0570 with the domain in header http://volker.dnsalias.net/ Please do not CC list postings to me.
Re: dodgy hd
On Tue, 2009-06-16 at 22:16 +1200, Volker Kuhlmann wrote: > On Tue 16 Jun 2009 21:27:10 NZST +1200, Steve Holdoway wrote: > > > > PS Did I say "smartmontools"? > > > > > Hmmm let's take this machine and have a look... > > The question was regarding a 120GB hard disk drive. There are none of > those that don't support smart. It became the "done thing" around 10GB > or so. Or in other words, there are no disk drives still under warranty > that don't do smart. The fundamental problem you're sidestepping is that there is no standard and no requirement for a monitoring interface on disks storage. To quote from the smartmon faq, "The raw SMART attributes (temperature, power-on lifetime, and so on) are stored in vendor-specific structures. Sometime these are strange." ( ain't that the case - smartctl reports one of my servers hdd temps between 60 and 65 degrees C - all 5 of them - when none of them are more than warm to the touch! But they're constant, except for slight seasonal changes which even seem to affect data centres worldwide ) Until there is a standard and requirement, you can't *guarantee* it will work or rely on the results. It'll *probably* work, there are plenty of cases where it plain doesn't work... like my SSD here - which is certainly under warranty, and, as you mentioned USB - which will become important soon when linux becomes the first OS to support USB 3. Until that time, smartmon is a really useful tool, but it's not the sole solution you make it out to be. Steve
Re: dodgy hd
On Tue 16 Jun 2009 21:27:10 NZST +1200, Steve Holdoway wrote: > > PS Did I say "smartmontools"? > > > Hmmm let's take this machine and have a look... The question was regarding a 120GB hard disk drive. There are none of those that don't support smart. It became the "done thing" around 10GB or so. Or in other words, there are no disk drives still under warranty that don't do smart. Of course if the disk is USB connected you're SOOL. Make appropriate temporary arrangements (any commercial tools won't save you here either because the USB adapter blocks the relevant low-level disk commands). Those mobile SATA racks are mightily convenient; for IDE, see SOOL above (-> open case, plug in cable to spare IDE connector). And if the disk is raid-controller connected, then it's basically 1) an Areca, and it just works 2) a 3ware, and support is pretty there (I understand) 3) some piece of rubbish, probably Highpoint, in which case it's "just desserts", 4) like 3, but you're using it in non-raid mode as simple controller, with Linux kernel raid, in which case smart should work too. So basically I don't see a problem, because there isn't one. If your hardware is buggered, you're wasting your time debugging your software, so smartmontools is mandatory, not optional. And it's step 1, not step 15. Volker -- Volker Kuhlmann is list0570 with the domain in header http://volker.dnsalias.net/ Please do not CC list postings to me.
Re: dodgy hd
On Tue, Jun 16, 2009 at 9:27 PM, Steve Holdoway wrote: > On Tue, 2009-06-16 at 20:45 +1200, Volker Kuhlmann wrote: > Warning: device does not support Error Logging > Error SMART Error Log Read failed > Smartctl: SMART Error Log Read Failed > Warning: device does not support Self Test Logging > Error SMART Error Self-Test Log Read failed > Smartctl: SMART Self Test Log Read Failed > Device does not support Selective Self Tests/Logging > > smartctl may be the answer... in a year or two. Not now. > > Its been a while since I saw a hard drive that didn't support smart reporting, so its likely a system misconfiguration. Exceptions being : a) Bizzare system harddrive controller that is only partially supported in linux b) Strange hard drive enclosures attached via USB so the required signals to transmit SMART data don't work. However, case b has also become less prevalent of late. -- Kent perl -e "print substr( \"edrgmaM SPA nocomil.i...@tfrken\", \$_ * 3, 3 ) for ( 9,8,0,7,1,6,5,4,3,2 );
Re: dodgy hd
On Tue, 2009-06-16 at 20:45 +1200, Volker Kuhlmann wrote: > > Any Linux user spending real dosh on M$-only products for analysing disks > instead of donating that to smartmontools does not deserve a free copy > of smartmontools. > > > Volker (amused) > > PS Did I say "smartmontools"? > Hmmm let's take this machine and have a look... Warning: device does not support Error Logging Error SMART Error Log Read failed Smartctl: SMART Error Log Read Failed Warning: device does not support Self Test Logging Error SMART Error Self-Test Log Read failed Smartctl: SMART Self Test Log Read Failed Device does not support Selective Self Tests/Logging smartctl may be the answer... in a year or two. Not now.
Re: dodgy hd
On Tue 16 Jun 2009 20:09:05 NZST +1200, Steve Holdoway wrote: > If you go through the possible reasons for the superblock to get > corrupted, then the causes are not limited to a hard disk failure. True. But then fsck would fix it *without* errors. In any case, smartctl will settle the issue. > Apart > from potential hardware issues elsewhere, there's configuration errors - > overlapping partitions, human error - dd of=/dev/sda just for a couple > of examples. None of that produces anything abnormal with smartctl. Nothing abnormal with smartmontools equals embarrassment when returning disk. > Sure, it's almost certainly a disk failure, but there are other possible > causes. Sure. How about just running smartctl instead of discussing possibilities ad infinitum? Either the disk or the data is foobared, in either case course of action is crystal clear. But hey, why not discuss it for another 3 days...? > Also, given the uphill fight you often get when not speaking > Microsoft to suppliers, A smartmontools printout showing the maker's own product says of itself(!!) "I'm stuffed on sectors XYZ" goes a long way. > As it happens, Dove are a pretty level headed supplier, and I've never > had problems returning failures. Yes yes yes I know. So go and take that damn disk back :) Any Linux user spending real dosh on M$-only products for analysing disks instead of donating that to smartmontools does not deserve a free copy of smartmontools. Volker (amused) PS Did I say "smartmontools"? -- Volker Kuhlmann is list0570 with the domain in header http://volker.dnsalias.net/ Please do not CC list postings to me.
Re: dodgy hd
On Tue, 2009-06-16 at 19:15 +1200, Volker Kuhlmann wrote: > I don't believe this thread goes on for that long. You buy a toaster > with 1y warranty and after 2 months half of it stops working. Is anyone > stupid enough to ask whether to return it?!?? > > Volker > If you go through the possible reasons for the superblock to get corrupted, then the causes are not limited to a hard disk failure. Apart from potential hardware issues elsewhere, there's configuration errors - overlapping partitions, human error - dd of=/dev/sda just for a couple of examples. Sure, it's almost certainly a disk failure, but there are other possible causes. Also, given the uphill fight you often get when not speaking Microsoft to suppliers, I think this thread is perfectly justified. As it happens, Dove are a pretty level headed supplier, and I've never had problems returning failures. Steve
Re: dodgy hd
On Tue 16 Jun 2009 13:51:12 NZST +1200, David Merriman wrote: > Something which I have found useful in some cases is SpinRite ( > www.grc.com/spinrite.htm ), which can perform low-level analysis and > recovery of hard drive sectors. It's not free, but has been > invaluable to me in the past (not recently, touch wood :) If it's for recovery, go ahead and spend money (if you were foolish enough to save on backups). If it's for analysis, why the hell would I waste my money on something the disk maker owes me for nothing? Either use the manufacturer's free tool, or run smartmontools. Besides, if the disk does have bad sectors within warranty you shouldn't have any problems getting it swapped. I don't believe this thread goes on for that long. You buy a toaster with 1y warranty and after 2 months half of it stops working. Is anyone stupid enough to ask whether to return it?!?? Volker -- Volker Kuhlmann is list0570 with the domain in header http://volker.dnsalias.net/ Please do not CC list postings to me.
Re: dodgy hd
2009/6/17 Barry Marchant > > > Steve Holdoway wrote: >> >> On Tue, 2009-06-16 at 07:19 +1200, Jim Cheetham wrote: >> >> >>> SAMSUNG HD753LJ s/n S13UJ1NQB01779 was the one that failed. Failure >>> was more than "just" the filesystem, I was unable to read or write to >>> the partition table. >> >> That comment was aimed at Barry, the OP who restored his filesystem >> using an alternate superblock and has had no problem since... >> > I have had no problems since, because I have done nothing with the faulty ptn > apart from copying everything on the ptn to a external hd. It has been > returned to Dove and I await their response. BTW the drive was a WD1600AAJB. > The superblock failed and running 'cp /dev/hda6 /dev/null' reported at least > 2 unreadable sectors. It still managed to boot on another ptn after I deleted > the faulty 1 from fstab, but with many complaints. > > THanks for all the replies > > Barry > Something which I have found useful in some cases is SpinRite ( www.grc.com/spinrite.htm ), which can perform low-level analysis and recovery of hard drive sectors. It's not free, but has been invaluable to me in the past (not recently, touch wood :) $0.02, David -- The meek don't want it.
Re: dodgy hd
Steve Holdoway wrote: On Tue, 2009-06-16 at 07:19 +1200, Jim Cheetham wrote: SAMSUNG HD753LJ s/n S13UJ1NQB01779 was the one that failed. Failure was more than "just" the filesystem, I was unable to read or write to the partition table. That comment was aimed at Barry, the OP who restored his filesystem using an alternate superblock and has had no problem since... I have had no problems since, because I have done nothing with the faulty ptn apart from copying everything on the ptn to a external hd. It has been returned to Dove and I await their response. BTW the drive was a WD1600AAJB. The superblock failed and running 'cp /dev/hda6 /dev/null' reported at least 2 unreadable sectors. It still managed to boot on another ptn after I deleted the faulty 1 from fstab, but with many complaints. THanks for all the replies Barry
Re: dodgy hd
On Tue, 2009-06-16 at 08:14 +1200, Jim Cheetham wrote: > Ah, OK. Down here in Dunedin I was probably too busy listening to the > list of school closures due to snow, and didn't check the comment > trail properly :-) > > -jim Heh - not here... yet (: Steve
Re: dodgy hd
On Tue, Jun 16, 2009 at 8:08 AM, Steve Holdoway wrote: > On Tue, 2009-06-16 at 07:19 +1200, Jim Cheetham wrote: >> SAMSUNG HD753LJ s/n S13UJ1NQB01779 was the one that failed. Failure >> was more than "just" the filesystem, I was unable to read or write to >> the partition table. > That comment was aimed at Barry, the OP who restored his filesystem > using an alternate superblock and has had no problem since... Ah, OK. Down here in Dunedin I was probably too busy listening to the list of school closures due to snow, and didn't check the comment trail properly :-) -jim
Re: dodgy hd
On Tue, 2009-06-16 at 07:19 +1200, Jim Cheetham wrote: > SAMSUNG HD753LJ s/n S13UJ1NQB01779 was the one that failed. Failure > was more than "just" the filesystem, I was unable to read or write to > the partition table. That comment was aimed at Barry, the OP who restored his filesystem using an alternate superblock and has had no problem since...
Re: dodgy hd
On Mon, Jun 15, 2009 at 10:15 PM, Hadley Rich wrote: > On Mon, 2009-06-15 at 22:08 +1200, Steve Holdoway wrote: >> Please tell me they weren't wd greens. I've had a 1TB fail within 5 >> minutes... in fact I've had more fail this year than in the last 10. > > Always interesting. I've had several quick Seagate failures lately, and > have been having a good run so far with the WD Greenpowers. SAMSUNG HD753LJ s/n S13UJ1NQB01779 was the one that failed. Failure was more than "just" the filesystem, I was unable to read or write to the partition table. Googling suggests http://www.samsung.com/global/business/hdd/productmodel.do?type=61&subtype=63&model_cd=248 -jim
Re: dodgy hd
On Mon, 2009-06-15 at 22:08 +1200, Steve Holdoway wrote: > Please tell me they weren't wd greens. I've had a 1TB fail within 5 > minutes... in fact I've had more fail this year than in the last 10. Always interesting. I've had several quick Seagate failures lately, and have been having a good run so far with the WD Greenpowers. hads -- http://nicegear.co.nz New Zealand's Open Source Hardware Supplier
Re: dodgy hd
On Mon, 2009-06-15 at 21:27 +1200, Jim Cheetham wrote: > On Mon, Jun 15, 2009 at 6:59 PM, Volker > Kuhlmann wrote: > > Any disk with bad sectors within warranty goes back to maker. Period. > > Agreed, I received two 750GBs about three months ago (with consecutive > serial numbers, sadly) and one has failed already, been straight back > to the vendor and been replaced. > Please tell me they weren't wd greens. I've had a 1TB fail within 5 minutes... in fact I've had more fail this year than in the last 10. It is questionable in this case whether it's a disk failure or not... it was the file system that fell apart and that's not *guaranteed* to be the disk at fault. Steve
Re: dodgy hd
On Mon, Jun 15, 2009 at 6:59 PM, Volker Kuhlmann wrote: > Any disk with bad sectors within warranty goes back to maker. Period. Agreed, I received two 750GBs about three months ago (with consecutive serial numbers, sadly) and one has failed already, been straight back to the vendor and been replaced. However, in my case diagnosis was easy; the drive was detected as being present by the OS, but nothing could be read from it at all (i.e. not even the partition table). Some vendors will try to reduce their return rate by saying that if they find nothing wrong, they will not replace the item. In these cases it's better to have a good set of diagnostic results available to show what you've done (i.e. SMART data, or possibly something from a liveCD like http://www.ultimatebootcd.com/). In any case, Dove should be reliable, tell them what's happened and organise a replacement drive. By default you need to return the old one first (one reason why I run everything on RAID1) but they may well be flexible if you talk to them. -jim
Re: dodgy hd
> oops, 120 gig from dove Any disk with bad sectors within warranty goes back to maker. Period. Volker -- Volker Kuhlmann is list0570 with the domain in header http://volker.dnsalias.net/ Please do not CC list postings to me.
Re: dodgy hd
oops, 120 gig from dove Craig Falconer wrote: Barry Marchant wrote, On 15/06/09 16:30: had 1 ptn on a 6 mth old 12G hd go belly up. 12 GB and 6 months old? Did you get it from a time warp or something?
Re: dodgy hd
Barry Marchant wrote, On 15/06/09 16:30: had 1 ptn on a 6 mth old 12G hd go belly up. 12 GB and 6 months old? Did you get it from a time warp or something? -- Craig Falconer
dodgy hd
Hi all, had 1 ptn on a 6 mth old 12G hd go belly up. Of course it was my main bootable ptn. All data was on other ptns and backed up so no major problems there. The crap ptn was recovered using fsck -b 32something and produced lots of inode an size problems. It now boots, I have not discovered anything missing yet, what further checks should I do, or should I just return it as being faulty? Barry