tag:blogger.com,1999:blog-4834634390856475978.post428531541299124636..comments2022-03-24T22:53:12.674+01:00Comments on Tim's Technical Thoughts: A note on Western Digital 2001FASS drivesTim Jacobshttp://www.blogger.com/profile/06131387085752434985noreply@blogger.comBlogger6125tag:blogger.com,1999:blog-4834634390856475978.post-17721628934799286422012-09-21T12:04:32.528+02:002012-09-21T12:04:32.528+02:00Update 2012: In the meantime, I've purchased a...Update 2012: In the meantime, I've purchased a Synology DS1812+ NAS, which accepts and works with the FASS2001 and FAEX2003 drives without any issues; in fact, I expanded from the original 5 2TB disks to 8 drives in a now fully packed Synology NAS. The Thecus has now been downgraded to a backup NAS which carries some (supported by Thecus) Seagate Barracuda 1 TB drives I still had lying around. Time solved the problem, as usual :).Tim Jacobshttps://www.blogger.com/profile/06131387085752434985noreply@blogger.comtag:blogger.com,1999:blog-4834634390856475978.post-56508942311761535052010-12-28T16:55:14.989+01:002010-12-28T16:55:14.989+01:00No, I have the full five disks running nicely agai...No, I have the full five disks running nicely again in RAID5 -- the disk with errors got fixed after a bad block scan; with bit error rates of 10^(-14) or once every 100 terrabits or roughly every 10 terrabytes, you can really expect bit issues in the lifecycle of a 2 TB disk (bit issue = the drive reads the opposite value of what was read due to magnetisation errors). To counter these bit errors, manufacturers include some "spare capacity" that you can use in case a bit error is encountered (=bad block) -- with TLER drives, this remapping of a bad block, this remapping is automatic I think, with the non-TLER drives it seems a bad block scan is necessary.<br /><br />I did that and now I still have the full capacity (2 TB) where now some bits are used from the extra spare capacity. Of course, the amount of spare capacity on a drive is finite, so once that runs out, the drive can no longer sustain its full 2 TB capacity due to too many bad block errors. That should trigger a SMART alert & allow the drive to be RMA'd (because vendors count that you don't hit the bad block count soon enough before the warranty expires). Of course, not having the full 2 TB capacity due to insufficient spare capacity to account for bad blocks, is bad in a RAID array where all drives are supposed to be created equally. <br /><br />Thusfar, I have not depleted the spare capacity, no SMART errors have been tripped so the drive is actually fully functional again according to Western Digital specifications. What concerns me is that the remapping of bad blocks to the spare capacity seems to require a manual intervention: removal of the drive (marked as failed by the Thecus NAS), bad block scan on another PC, then remapping bad blocks to spare drive capacity and finally reinserting in the NAS and rebuilding the RAID array. That should be automatic and that should be the case with the WD RE series of drives.<br /><br />Also, I don't blame Thecus for not supporting the drive -- they are in fact correct to state that this is not a supported drive, simply due to the fact that the drive firmware is CRIPPLED such that it does not seem to do bad block remapping automatically (or fast enough). What I don't understand is that drives such as the RE4 which support TLR are not on the Thecus compatibility list. So to conclude: the blame is with WD, not Thecus (at least IMHO).<br /><br />That being said, my FASS2001 drivers are working perfectly fine again now -- I'm just a bit pissed because of the entire TLER/crippled firmware/expensive drives situation which still seems to require a manual intervention from my part in case of RAID troubles (which is precisely why I coughed up €800 for a Thecus NAS -- to be 100% sure that when things go wrong, I still have my data!!).Tim Jacobshttps://www.blogger.com/profile/06131387085752434985noreply@blogger.comtag:blogger.com,1999:blog-4834634390856475978.post-1346926390026463032010-12-28T16:55:04.571+01:002010-12-28T16:55:04.571+01:00No, I have the full five disks running nicely agai...No, I have the full five disks running nicely again in RAID5 -- the disk with errors got fixed after a bad block scan; with bit error rates of 10^(-14) or once every 100 terrabits or roughly every 10 terrabytes, you can really expect bit issues in the lifecycle of a 2 TB disk (bit issue = the drive reads the opposite value of what was read due to magnetisation errors). To counter these bit errors, manufacturers include some "spare capacity" that you can use in case a bit error is encountered (=bad block) -- with TLER drives, this remapping of a bad block, this remapping is automatic I think, with the non-TLER drives it seems a bad block scan is necessary.<br /><br />I did that and now I still have the full capacity (2 TB) where now some bits are used from the extra spare capacity. Of course, the amount of spare capacity on a drive is finite, so once that runs out, the drive can no longer sustain its full 2 TB capacity due to too many bad block errors. That should trigger a SMART alert & allow the drive to be RMA'd (because vendors count that you don't hit the bad block count soon enough before the warranty expires). Of course, not having the full 2 TB capacity due to insufficient spare capacity to account for bad blocks, is bad in a RAID array where all drives are supposed to be created equally. <br /><br />Thusfar, I have not depleted the spare capacity, no SMART errors have been tripped so the drive is actually fully functional again according to Western Digital specifications. What concerns me is that the remapping of bad blocks to the spare capacity seems to require a manual intervention: removal of the drive (marked as failed by the Thecus NAS), bad block scan on another PC, then remapping bad blocks to spare drive capacity and finally reinserting in the NAS and rebuilding the RAID array. That should be automatic and that should be the case with the WD RE series of drives.<br /><br />Also, I don't blame Thecus for not supporting the drive -- they are in fact correct to state that this is not a supported drive, simply due to the fact that the drive firmware is CRIPPLED such that it does not seem to do bad block remapping automatically (or fast enough). What I don't understand is that drives such as the RE4 which support TLR are not on the Thecus compatibility list. So to conclude: the blame is with WD, not Thecus (at least IMHO).<br /><br />That being said, my FASS2001 drivers are working perfectly fine again now -- I'm just a bit pissed because of the entire TLER/crippled firmware/expensive drives situation which still seems to require a manual intervention from my part in case of RAID troubles (which is precisely why I coughed up €800 for a Thecus NAS -- to be 100% sure that when things go wrong, I still have my data!!).Tim Jacobshttps://www.blogger.com/profile/06131387085752434985noreply@blogger.comtag:blogger.com,1999:blog-4834634390856475978.post-46974548806864616062010-12-28T15:47:43.205+01:002010-12-28T15:47:43.205+01:00So you are now running with 4 disks in degraded mo...So you are now running with 4 disks in degraded mode..?<br />Brave man :)<br /><br />I saw on the Thecus site that they already covered their ass by saying WD disks might have this issue<br />http://www.thecus.com/Downloads/HDD_List/N7700_N7700SAS_N8800_N8800SAS_SATA_HDD_list_2010-12-10.pdf<br /><br />[Notes for WD 1.5/2.0 TB HDD]<br />1. WD2002FYPS, WD20EADS, and WD15EADS may have compatibility issue with Thecus<br />N7700/N7700SAS/N8800/N8800SAS<br />2. About 90% of WD2002FYPS/WD20EADS/WD15EADS configuration is stable, but remaining<br />10% might experience RAID degrade issue randomly.<br />SO those 10% are getting no love from Thecus..good luck and keep us posted on what adventures you encounter in the future!Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-4834634390856475978.post-68465725038162143732010-12-28T13:23:06.947+01:002010-12-28T13:23:06.947+01:00Up until now, I haven't bought any additional ...Up until now, I haven't bought any additional harddisks. Perhaps that I will eventually cough up the extra bucks for the RE version of the Western Digital drives. <br /><br />I didn't return the drive since a full bad block scan "fixed" the drive (reallocated bad sectors). Of course, when the spare capacity for the bad sectors runs out, I will return it and use my 5 year guarantee. <br /><br />I'm not sure what Western Digital will say when the drive is RMA'd with the explanation that it doesn't work in a RAID array. In fact, the drive is advertised as not being able to do that :). I suppose that, assuming they actually read the explanation before swapping the drive, they might refuse an RMA. But then again, I will not send it back unless it has really failed.<br /><br />In the meantime, I can say that my Thecus is happily up and running again without any issues ever since.Tim Jacobshttps://www.blogger.com/profile/06131387085752434985noreply@blogger.comtag:blogger.com,1999:blog-4834634390856475978.post-40453105014618932612010-12-28T11:32:24.964+01:002010-12-28T11:32:24.964+01:00What did you end up buying instead?
Where you able...What did you end up buying instead?<br />Where you able to return the drive?<br />If you return the drive with the explanation that your NAS keeps detecting it as faulty what can they say? (besides see your NAS supplier :) )<br /><br />This is becoming a problem with probably most if not all HD suppliers. They consider the TLER function to be a "PRO" feature while more and more consumers are purchasing RAID cards/NAS/ etc for home use.<br />Great story BTW!Anonymousnoreply@blogger.com