Need advice to combat warranty monkey-business by LaCie (Seagate)

[sbpetrack]

I have a La Cie P9227 Slim 2TB usb 3.0 drive; it's a slim aluminum enclosure which contains a Seagate ST2000LM007-1R8174 2000.3GB drive. One fine day, programs began to freeze up at random moments when the disk was accessed. So the first thing I did of course was try to read all data that did not have a current back-up. This took several days but in the end i prevailed, so now I have no need of any of the data on the drive.
Next I ran SeaTools for Windows on the drive, hoping either to fix the problem or to get a clear fail which would let me RMA the drive. The SeaTools DST (drive self test) passed, and the Short Generic Test ran until about 25% and then didn't advance for two days, at which point I abandoned it. Hoping that "all" i needed to do was reallocate bad sectors, I decided to take the drive out of the enclosure, attach it internally via SATA, and run chkdsk /r. Unlike SeaTools, chkdsk gives an indication of just which sector it's working on. I could see it become INSANELY slow at the 25% -- but it didn't stop. 24 hours later, it got to 62%.... where it stayed on the same sector for 2 days, at which point i killed chkdsk.
I still had neither a clear Fail i could show LaCie, nor a repaired disk I could use to get on with my life. But since chkdsk had now repaired the damage around the 25% mark, and the disk was attached via SATA, I decided to try the SeaTools "Fix All Long" test. This took only about 10 hours to get to the 62% and then... FAIL.
At this point, I contacted LaCie, and got a polite response which thanked me for having done so much to try to repair the drive.... and then ended with an offhand remark that by opening the enclosure I had voided the warrantee, so i could continue however i wanted to. "thanks for choosing LaCie".
I have since read the manual of the drive and the warrantee very carefully, and it mentions NOTHING about such a clause. And on the contrary the manual of SeaTools makes it VERY clear that to improve both the quality and speed of the diagnosis and repair, you should attach the drive via SATA rather than with a USB controller. Of course it does state that they are not liable for any damage caused by abuse or user-inflicted mishandling, but there is none in this case. I was very careful to use the proper nylon scribe to open the case, since it was one of these "Porsche Design" things that had been a gift. I would have no problem to accept that they will not replace the aluminum Porsche-designed enclosure btw, if they want to claim that merely opening it voids the warrantee for IT. It's the disk that has a problem, not the case. They pointed me to a page on their web-site with a warrantee-void checklist. But quite frankly the manual makes no mention of this page, and on the contrary says explicitly that the user should read very carefully the terms and conditions of using the drive and the warrantee that are in the manual -- so i can't see how i could been expected to read some random footnote on their web site to know that I shouldn't have opened the case.
It burns me up that I took a LOT of time in order to do what i could to repair the drive, to be rewarded in this fashion. I have replied to LaCie tech support, outlining BOTH the simple legal case as I've stated it here, and also asked them that EVEN if somehow I missed something and indeed they could legally refuse to replace the drive, they will surely be able to tell that no physical damage was caused by opening the case, and perhaps they could escalate a request to make an exception to their hidden policy. To my surprise, I have not received a reply, at least not yet.
I would just like to ask people here, the ones who have a great deal more experience than I with LaCie/Seagate and similar companies, if 1) it is just 'obvious' and generally understood that taking a drive out of its USB enclosure voids the warrantee of the drive inside, and 2) if there is some simple escalation path available to me, should I continue to receive no reply, or receive a refusal to exchange the drive.
This is purely a legal/procedural question about getting a company to honour its guarantee. I'll ask in a different thread if there's a FAST way to get those bad sectors reallocated or hidden .

Thanks, scott

 

[Supahos]

It's not monkey business. I bet you even cut a warranty sticker to get it open. You should have contacted them a long time before you popped open the drive.

 

[gdmaclew]

Supahos is right. Warranties are an "either-or" situation. Either you get a replacement under warranty OR you try to diagnose/fix it your self. You can't do both.

 

[USAFRet]

I decided to take the drive out of the enclosure, attach it internally via SATA

And that killed the warranty.

If it was throwing up errors, your first move should have been to contact LaCie.

 

[gdmaclew]

MERGED QUESTION
Question from sbpetrack : "Is there a FAST way to deal with a Current Pending / Uncorrectable Sector Count? a FASTEST way?"

I have a Seagate 2TB SATA HDD which seems to have decided that some of its sectors are an ideal place for data to start playing Hide&Seek. I was able to restore or recover all data; then I formatted the disk as a single large NTFS partition, so that I could run chkdsk /r. This ran until it got to a certain sector at about 62% (stage 5). Two days later it was still on the same sector, so i killed it. I then tried the "Fix All Long" uitility of SeaTools for WIndows. This took about 10 hours to get to the same 62% and then... FAIL. I took a look with CrystalDiskInfo:

I think that this means that there are 312 (= 0x138) uncorrectable sectors and they are all still considered to have data in them by Windows. But I already know that I don't need any data on this disk.

I understood from this forum that if I just write zeros to every sector then bad sectors will be reallocated. But I assume that if I just do "dd" for example, the write will hang at exactly the same 62% mark where it stopped already twice.

Among all the various disk and other utilities, is there one that has the best chance to remove any bad segments from use, in the most robust and quickest manner, and which is least likely to get stuck on some sector? I don't care if it works by writing zeros or by some other magic.

If there is a trade-off to be had, i am perfectly willing to lose forever some sectors that a slower approach would have managed to save.

As an example of something that is not a good solution here, I did start to run SprinRite Level 4 on the disk. When its estimate of how much time it will need to complete the disk reached 1500Hours, I killed it. Or is it reasonable to run SpinRite for 62.5 days to solve this problem?

Replace the drive.

Replace the drive.
You will get the same answer to both of your posts on this same problem.

 

[sbpetrack]

well live and learn. I certainly won't argue with you, but I have to mention that certainly the LaCie web-site go out of their way to encourage a user to do as much self-help as possible, and only to contact them when the repair tools have been exhausted. And just for the record, I certainly didn't cut any warranty sticker. Nor is there anything on the model/serial number label (which is the only writing on the disk) which says that opening the enclosure voids the warranty. From my point of view, this is a textbook example of an old Yiddish saying: "no good deed goes unpunished."

Since my two questions have been merged, i'll just mention that after writing my question about how to deal with the bad sectors, I thought "why not try dd, indeed?" Could it be that simple? I use the excellent Windows rawwrite dd for lots of things (it's at http://www.chrysocome.net/dd), maybe it's also a harddisk utility?!? I entered the following command at an elevated command prompt:

dd if=/dev/zero of=\\?\Device\Harddisk3\Partition0 seek=1 bs=50M --progress

(The seek=1 is there in case it's a bad idea to zero out the very first bytes of the disk. The fact is that right now the disk is completely unallocated. I just didn't know if there should be a correct but "empty" MBR, for example, or if one can just write 0s from beginning to end. And although the blocksize of 50M is certainly unorthodox, this machine has 20G of RAM, so it can handle it. As I write this, dd is busy writing 0s in 50M blocks to this drive at rate which - if it continues --will give me my disk back in 4-5 hours. I would normally wait before posting this -- I do so not to brag in advance, but just in case someone knows already that this will be a terrible idea when it hits the damaged sectors, which it will do in about 45 minutes. If so, i would be grateful for the heads-up.

 

[gdmaclew]

"Self-help" does not include getting out your screwdriver and opening up a product. The drive comes in a case and that's part of the product. Disassembly of any kind voids the warranty and I'm 100% sure Lacie does NOT suggest this.

 

[sbpetrack]

first of all, no screwdriver was involved, it was a nylon scribe. Second, the SeaTools user manual states explicitly things like "fix all fast" and "fix all long" 1. work only on Seagate drives; 2. require that the disk be attached via SATA; and 3. provide better diagnostics, more comprehensive restoration, AND do it faster than what the long generic test can offer (which is the only routine available to a USB-attached disk ).
Now indeed one needs to be an adult here. If you are a user and looking at a sealed enclosure with a USB port, there's a good chance you're not even aware of the possibilty that better repair is possible. But I respectfully assert that 1. YES, if you are competent to take out the drive and have enough knowledge to realize that the disk inside is likely to be SATA, then **OF COURSE** you will read those facts and you'll take out the drive. In fact, I'll go so far as to say that given what i have now learned from you, I think it's misleading and irresponsible for the SeaTools user manual NOT to include a sentence among the ones that explain the advantages of running the SATA-based test -- a sentence which asserts VERY CLEARLY something like "WARNING! REMOVING A DRIVE FROM AN ENCLOSURE FOR THE PURPOSE OF RUNNING ONE OF THE FIX ALL TESTS WILL ALMOST CERTAINLY CAUSE ANY WARRANTY TO BECOME VOID IMMEDIATELY." Because yes, Seagate *is* essentially encouraging the user to take out the drive.

In my case, the enclosure was actually rather tricky to open, and i'll admit that there was a moment when i knew i was doing something which had some risk. The risk I **thought** i had was the risk that if i slipped and damaged the enclosure, i might void the warrantee. ON THE ENCLOSURE. I doubted very much i was going to damage the disk -- though that was possible, true. And if i *had* slipped and dropped or whacked the disk -- then of COURSE i would have to suffer the consequent loss of the warranty. But I didn't -- because i didn't DO anything but open the enclosure and take out the disk. BUT WAIT!!! THere's MORE

About 3 hours after i hit <enter> on the dd command above, dd returned with the following error:

C:\WINDOWS\system32>dd if=/dev/zero of=\\?\Device\Harddisk3\Partition0 seek=8 bs=50M --progress
rawwrite dd for windows version 0.6beta3.
Written by John Newbigin <jn@it.swin.edu.au>
This program is covered by terms of the GPL Version 2.

1,622,600M Error writing file: 2 The system cannot find the file specified
1,622,600M
32453+0 records in
32452+0 records out

I noted with a mixture of optimism and concern that 1,622,600M is just over 85% of the total capacity of the drive, which reported by to be 2000398934016 bytes. So dd.exe had succeeded to write 0s past BOTH the 25% mark AND the 62% limit -- the latter limit having stopped both chkdsk and the long generic test of SeaTools. And the whole thing took less than 4 hours!! The failed tests had taken almost 21 hours to fail, and others had gone of for two days before i cracked and aborted. And then I ran CrystalDiskInfo on the drive:

Does this mean that my drive is now actually (in principle) COMPLETELY fixed? IN fact, not only are all fields ''good", I note that the Reallocated sectors count is still 0!! Do I understand from this that ALL 0x138 uncorrectable sectors from the original parameters in the end just corrected themselves? Do I need to worry about the huge increase in the Head Flying Hours?

Remembering that SpinRite said it would need 1500 hours to scan the drive and refresh the defect list, I feel that something is fishy about a solution that takes 3-4 hours. Should I be worried about the health of the disk? Or can I rely on the SMART data to tell me that I have a 'new' drive. THanks!

 

[aquielisunari]

first of all, no screwdriver was involved, it was a nylon scribe. Second, the SeaTools user manual states explicitly things like "fix all fast" and "fix all long" 1. work only on Seagate drives; 2. require that the disk be attached via SATA; and 3. provide better diagnostics, more comprehensive restoration, AND do it faster than what the long generic test can offer (which is the only routine available to a USB-attached disk ).
Now indeed one needs to be an adult here. If you are a user and looking at a sealed enclosure with a USB port, there's a good chance you're not even aware of the possibilty that better repair is possible. But I respectfully assert that 1. YES, if you are competent to take out the drive and have enough knowledge to realize that the disk inside is likely to be SATA, then **OF COURSE** you will read those facts and you'll take out the drive. In fact, I'll go so far as to say that given what i have now learned from you, I think it's misleading and irresponsible for the SeaTools user manual NOT to include a sentence among the ones that explain the advantages of running the SATA-based test -- a sentence which asserts VERY CLEARLY something like "WARNING! REMOVING A DRIVE FROM AN ENCLOSURE FOR THE PURPOSE OF RUNNING ONE OF THE FIX ALL TESTS WILL ALMOST CERTAINLY CAUSE ANY WARRANTY TO BECOME VOID IMMEDIATELY." Because yes, Seagate *is* essentially encouraging the user to take out the drive.

In my case, the enclosure was actually rather tricky to open, and i'll admit that there was a moment when i knew i was doing something which had some risk. The risk I **thought** i had was the risk that if i slipped and damaged the enclosure, i might void the warrantee. ON THE ENCLOSURE. I doubted very much i was going to damage the disk -- though that was possible, true. And if i *had* slipped and dropped or whacked the disk -- then of COURSE i would have to suffer the consequent loss of the warranty. But I didn't -- because i didn't DO anything but open the enclosure and take out the disk. BUT WAIT!!! THere's MORE

About 3 hours after i hit <enter> on the dd command above, dd returned with the following error:

C:\WINDOWS\system32>dd if=/dev/zero of=\\?\Device\Harddisk3\Partition0 seek=8 bs=50M --progress
rawwrite dd for windows version 0.6beta3.
Written by John Newbigin <jn@it.swin.edu.au>
This program is covered by terms of the GPL Version 2.

1,622,600M Error writing file: 2 The system cannot find the file specified
1,622,600M
32453+0 records in
32452+0 records out

I noted with a mixture of optimism and concern that 1,622,600M is just over 85% of the total capacity of the drive, which reported by to be 2000398934016 bytes. So dd.exe had succeeded to write 0s past BOTH the 25% mark AND the 62% limit -- the latter limit having stopped both chkdsk and the long generic test of SeaTools. And the whole thing took less than 4 hours!! The failed tests had taken almost 21 hours to fail, and others had gone of for two days before i cracked and aborted. And then I ran CrystalDiskInfo on the drive:

Does this mean that my drive is now actually (in principle) COMPLETELY fixed? IN fact, not only are all fields ''good", I note that the Reallocated sectors count is still 0!! Do I understand from this that ALL 0x138 uncorrectable sectors from the original parameters in the end just corrected themselves? Do I need to worry about the huge increase in the Head Flying Hours?

Remembering that SpinRite said it would need 1500 hours to scan the drive and refresh the defect list, I feel that something is fishy about a solution that takes 3-4 hours. Should I be worried about the health of the disk? Or can I rely on the SMART data to tell me that I have a 'new' drive. THanks!

First off La Cie is not guilty of any monkey business. You abused their product. You removed the enclosure that should only be removed by an authorized rep..

Secondly IDK. I would never use that drive for my main OS or for anything important.

 

[sbpetrack]

Just for those people who might be interested in either of the two separate issues, here is an update:

Warranty: I simply wrote back to tech support, pointing out the facts that to me make an unassailably strong case: 1. the manual makes no mention whatsoever about opening the enclosure and it **does** explicitly forbid opening the drive. 2. The manual points the user to SeaTools, and the SeaTools documentation states very explicitly (on page 8) that if the short drive self test fails, then a supported Seagate SATA drive should use the one of the "fix all" tests while a USB drive should use the "long generic" test. So I ran the long generic test.... and it failed too. I then did what seemed and still seems to me to be the obvious next step: I looked everywhere to see if the drive itself is a "supported Seagate SATA drive" -- but there is no such list. So I took the drive out to run the "fix all long" test. ( I wish to state for the 'experts' reading this that BEFORE I ran this 'fix all long' test, I spent several hours in a last attempt to recover data from the drive. And I succeeded. Because the drive was now SATA connected, I had available some recovery tools that weren't available for the USB-connected drive. Not to mention how much faster they worked. Anyone who says I needed to return the drive without taking it out of the enclosure is saying that if I wanted to exercise my legal and moral right to get a replacement drive for this broken one. then I would have to accept data loss, because La Cie would not have been willing to recover any data from the drive. I'm not sure this attitude qualifies as customer-centric or even as reasonable. If you want to be hard about it, then say if the customer damages, abuses, or even just opens the case, then the case is no longer covered by warranty. Follow this sentence with the same sentence where the word "case" is replaced by "hard drive." I personally find it somewhat offensive that a company's warranty policy should force me to choose between recovering my data and getting a replacement for a faulty drive; it is certainly unnecessary. )
Having got my data, I ran the 'fix all long' test -- and it failed. It's at this point that I contacted LaCie about a replacement drive. And when they told me that the warranty was now void, I replied to them with the facts as I just stated them. And to be extra courteous, I offered to exchange only the hard drive itself, and to keep the case.

I am now pleased to report that the reply I got back had two parts:
1.i was pointed to what they said was the relevant sentence in the warranty: which is that the warranty is void if I use the product "in a manner not intended by LaCie." ( I have since replied to them that in this case, I am entirely vindicated, because SeaTools makes it 100% clear that Seagate intends you to use the "fix all long" test on Seagate supported drives, and since there isn't any list of such anywhere, the only possible way to proceed is to try it. That LaCie didn't "intend" me to take the drive out of the case is a notion so ridiculous as to be laughable. I'm sure they didn't intend for me to use a drive with bad sectors in it either. Does that mean that the warranty is void if the there are bad sectors on the drive? Not only that, but there are at least two industry-standard ways that a company can make its "intentions" known on these matters: they can put warranty sticker or some equivalent mechanism to explicitly void the warranty, and/or they can add a sentence to their documentation. The manual clearly states that by using the product I agree to abide by the various terms and conditions. I checked, and nowhere do the Ts & Cs say that I need to be a mind-reader of LaCie's implicit intentions. )
2. They graciously agreed to replace the drive, on condition that I send them first some pictures of the reassembled drive which will prove to them that the case is not damaged. I of course agreed immediately, and sent them some pics immediately. Although I couldn't help myself but note that they should be more concerned about damage I might have done to the DRIVE than to the case.

Now without any hint or intention of being snide or cynical, I say that I gave all these details in case indeed someone reading this is responsible for crafting the RMA policy of a drive manufacturer. IMHO, as long as there are recovery possibilities available to the drive-outside-the-enclosure, it is just not right to consider opening the case to be ipso facto abuse. And it is particularly egregious to consider it abuse but not to say so explicitly. And triply so to consider it abuse, not to say so explicitly, and to offer your users recovery software which has recovery features that can ONLY be used on hard drives that are outside the case, and which MUST be run as part of the warranty-return service. That's just Kafkaesque.

epilogue: It was only a few hours after I replied to LaCie that I had the bright idea just to use dd to write zeros to the drive as a way to update the list of grown defects. The next thing I did was to format the disk as one big NTFS partition (this was a slow format, not a quick format). After I did this, I looked at the SMART data again, and now, to my amazement, it reported 8 reallocated sectors. I then once again ran the long generic test on the drive (back in its enclosure). It took 23 hours to complete. And it Passed!! It has been a few days since then, and Crystal Disk Info tells me that the SMART data has not changed. So whatever happened, it caused 8 sectors to go bad and be reallocated.

If anyone has read this far (!) and can give me some clue as to the size of the pool of spare reserved sectors I could expect from a 2.5 inch 2TB HDD, I'd be grateful to know. In the meantime, I am waiting to hear from LaCie to know if we should continue with the RMA or not.

-scott