1.5TB Barracuda freezes: Seagate speaks out

Last week, we stumbled upon multiple user reports of intermittent hangs with Seagate’s 1.5TB Barracuda 7200.11 hard drive. According to folks on the Seagate support forums, the drives randomly freeze for around 30 seconds at a time in Linux, Mac OS X, and Windows Vista. Freezes allegedly occur in tasks like streaming video or transferring files at low speeds, and they happen regardless of drive configuration or Serial ATA controller type.

We shot off an e-mail to Seagate last week to inquire about these problems. We’ve now received a response from company spokesman Mike Hall, who’s written the following:

Seagate is investigating an issue where a small number of Barracuda 7200.11 (1.5TB SATA) hard drives randomly pause or hang for up to several seconds during certain write operations. This does not result in data loss nor does it impact the reliability of the drive but is an inconvenience to the user that we are working to resolve with an upgradeable firmware.
We are therefore asking customers if they feel they are experiencing this issue to give our technical support department a call with any questions.

Affected part number: 9JU138-300, 336 with firmware revisions SD15, SD17, or SD18.

Support contact information:

Technical Support in the USA: 1.800.SEAGATE (1.800.732.4283)

Technical Support in Canada: 1.405.324.4700

Other regions please go to our support web page: http://www.seagate.com/www/en-us/support/

For the record, many users who posted in the Seagate support forums say they’ve received unhelpful responses from the company’s support staffers, including some saying the 1.5TB ‘cuda doesn’t support Linux or RAID setups. (Desktop RAID is still among the list of “best-fit applications” on the official product page.) One user posted about a similar response as early as last Saturday.

Comments closed
    • digiteyes
    • 12 years ago

    We have also experienced problems with the 1.5 TB Seagate drives here at work, specifically with Quicktime movies. We buy the drives in batches of 20 since we routinely work with 150 GB quicktime files. We have not experienced problems with any of the newer models but we do have some with the SD17 firmware that cause problems, even though we used Seagate’s online utility and it claims our serial number for the suspect drives should be in the clear… they are not.

    We specifically see problems when saving out self-contained Quicktime movies from a Quicktime reference file. When we first generate a file from Final Cut Pro, we save them as Quicktime Reference movies because it is so much faster. Then when we are ready for final output, we save all of those reference movies (typically overnight) as self-contained movies. If the reference files are on one of the SD17 drives, the save-out process stalls and then the entire Mac must be restarted. Usually, it will require a “force restart” because just trying to do a restart or shutdown from the menu will have it sitting there doing nothing for hours.

    Anyhow, confirmed again today that it was, in fact, the Seagate firmware bug by copying all of the reference files that kept freezing to my internal Western Digital drive. They saved out with no problems from the WD drive, when I had spent all day yesterday failing from the Seagate (it wasn’t until the end of the day that my boss suggested it might be the Seagate Firmware bug – good guess).

    Not bashing Seagate (though I hate that they are being a bit shifty and hush-hush about this problem) since, like I said, all of the newer drives we have from them work well. If anything, for our specific problem I would be more inclined to blame Apple for the way Quicktime generates data – a pause or interrupt should be easy to get past. It should just pickup writing the file where it left off if an interrupt occurs; other file types have been able to do this for years.

    • kick
    • 12 years ago

    Last year I buy 2 HDD Seagate 640 Gb ST 3640323AS I use them in Raid 0 Mainboard Gigabyte X48T-DQ6 ICH Controler, Firmware SD13 operating system Vista Home Basic x64 and from time to time Vista was freezing I can move only the mouse and after some time stop responding at all.

    This year I buy 2 HDD Seagate 1Tb ST31000333AS same computer same OS same RAID same drivers, Firmware CC1H and are making same problem. I ask Gigabyte support and they send me to Seagate it’s looking that the problem from here it’s coming and from all 7200.11 hard drives. I found more messages reporting that 1.5 Tb its making this problems but for me it’s looking that all has some problems hardware or firmware.

    Dear Seagate do you have any suggestion?

    What to do with these HDD?

    • gunga64
    • 12 years ago

    Well I have had a ton of issues with seagate even using 750 GB drive raids. The 1 TB drives are even worse. They freeze up randomly while copying then start back up slowly. Also these freezes dont have anything to do with them going into sleep mode as they just happen while already in write state. I do think turning the cache off helps. But I dont want to keep doing that.

    I have switched back to wd for sata and have had better luck. It seems if you do do not use raid or use these drives in an external usb enlosure the drives work fine. But in both cases I have found these drives to be highly unreliable. Constantly falling out of raid or disconnecting from a usb connection.

    Also I’d think twice if you think seagate will make firmware to fix these problems. I have called seagate many times in the past for firmware only to be told there are no firmware patches for my drives. When I know there is! Sorry seagate you are dropping the ball here. The westerndigital re3 drives seem to be more stable.

    I have just about given up on seagate but was going to look into them for 1.5Tb as western digital does not have one yet. But this even sounds worse then the 750 gb crashes. Also one last thing. These drives seem to get back into raid again after a restart of the system. But it would break the raid in a few weeks. Even after swapping out new drives. Finally many of the drives just failed to work at all. So they do fail period at the end. Seagate seems to think there is not a problem for the most part.

    • bend0r
    • 12 years ago

    The fact that Seagate made millions of these drives and only a select few are having issues automatically makes the drives bad and Seagate at fault?

    Good reasoning there. I have one of these drives right now and it works flawlessly. Since the problem seems to be in a RAID, how about checking with the RAID controller manufacturer? Nvidia saw this issue early and release a new driver to fix the issue. AMAZAZING.

    Desktop RAID = RAID 0 or 1, not RAID 5, go get one of their ES drives for that.

      • tenortim
      • 12 years ago

      “Desktop RAID = RAID 0 or 1, not RAID 5, go get one of their ES drives for that.”

      Really? Care to back that unfounded assertion up with ANY kind of factual basis? Like anything on the Seagate web site, or anywhere else for that matter?

      I built a system several months ago for my wife. She does a lot of PhotoShop work. It’s a regular Intel desktop system (P35 chipset) and, since it has an ICH10R as the South bridge, it supports RAID5 and her disk setup is indeed a RAID 5 setup to allow for lots of room for the large RAW image files she works with. In what way is this not “desktop RAID”?

      If the drive only supports RAID 0 and RAID 1 then that should be clearly and unequivocally stated. It would, however be be ridiculous – you’d still be subject to the same timeout issues, and if it were RAID 0, you’d still be screwed.

      The reason people are angry is not because the product is flawed, but the lies and evasions coming out of Seagate. How about “Yes, we screwed up. Yes a lot of drives are affected. Yes we are working on it. Yes, we will make sure that this gets fixed and you can get the firmware upgraded. Sorry”. Rather than, “(initially) no it’s not a problem… (subsequently) well uhm, yes, but hardly anybody is affected, and why would you be using it for something that our data sheet said you could use it for anyway?, oh, you’re using Linux/BSD/MacOS/Vista/anything else, oh well, it isn’t supported for that!”

      • fluke
      • 12 years ago

      “The fact that Seagate made millions of these drives and only a select few are having issues automatically makes the drives bad and Seagate at fault?”

      No, the fact that the same exact problem can be reproduced across multiple drives in the same production batch and Seagate fails to provide replacements is what makes Seagate at fault. This isn’t an issue of a drive “automatically” being bad. This is an issue of lack of warranty support.

      “I have one of these drives right now and it works flawlessly.”

      You just have a Barracuda 7200.11 or you have one of the ones listed with the affected part number? If you survived the Seagate roulette then I’m happy for you. But a warranty isn’t for when everything works flawlessly. The warranty is supposed to protect you for when things don’t work flawlessly. Lack of good warranty support can quickly become a major bad reputation issue. IBM Deskstar product became dubbed the “Deathstar” because of high failure rate. The problem stated at first with a select few drives for which IBM failed to acknowledge any warranty issue. I know people that still refer to the Hitachi line of drives the same way despite it coming from a different company now.

      The question shouldn’t be does a drive automatically become bad but should an entire vendor automatically be considered becoming bad. Can we count on Seagate to honor their warranty? What printed on Seagate’s website acknowledges the policy their own support representatives have stated that drive replacement can not be performed due to use of GNU/Linux? What does it mean when Seagate’s website says that a drive is good for “Desktop RAID” but support states it is “not meant to be used in a RAID environment.” What of Seagate’s documentation or advertising material can be trusted to be accurate (as in actually a statement that support honors) and what is as useful a claim as if it came from an used car saleman?

      “”The fact that Seagate made millions of these drives and only a select few are having issues automatically makes the drives bad and Seagate at fault?”

      No, the fact that the same exact problem can be reproduced across multiple drives in the same production batch and Seagate fails to provide replacements is what makes Seagate at fault. This isn’t an issue of a drive “automatically” being bad. This is an issue of lack of warranty support.

      “I have one of these drives right now and it works flawlessly.”

      You just have a Barracuda 7200.11 or you have one of the ones listed with the affected part number? If you survived the Seagate roulette then I’m happy for you. But a warranty isn’t for when everything works flawlessly. It is supposed to protect you for when things don’t work flawlessly. Lack of good warranty support can quickly become a major bad reputation issue. IBM Deskstar product became dubbed the “Deathstar” because of high failure rate. The problem stated at first with a select few drives for which IBM failed to acknowledge any warranty issue. I know people that still refer to the Hitachi line of drives the same way despite it coming from a different company now.

      The question shouldn’t be does a drive automatically become bad but should an entire vendor automatically be considered becoming bad. Can we count on Seagate to honor their warranty? What printed on Seagate’s website acknowledges the policy their own support representatives have stated that drive replacement can not be performed due to use of GNU/Linux? What does it mean when Seagate’s website says that a drive is good for “Desktop RAID” but support states it is “not meant to be used in a RAID environment.” What of Seagate’s documentation or advertising material can be trusted to be accurate (as in actually a statement that support honors) and what is as useful a claim as if it came from an used car saleman?

      “Since the problem seems to be in a RAID, how about checking with the RAID controller manufacturer?”

      I have contacted 3ware, LSI and Adaptec and all three are looking into the issue. But should it really be considered multiple RAID controller manufacturer’s responsibility to fix an issue with a single line of hard drives which does not even meet it’s own stated specifications? How can a drive have a specification of a random read seek time of < 8.5 msec and then support indicate there is no problem with a pair of drives that will periodically be unresponsive for over 100 msecs? I can accept that there may be a problem where one drive periodically has problems. I can accept that there might be a line of RAID controller that doesn’t work. But when BOTH drives get marked as faulty by a 3ware 8006, 3ware 9650SE, LSI 1068E and Adaptec 5405 then it seems like the drives really are faulty. Or do you believe that all of these vendors have bad firmware or drivers?

      The bottom line is these drives do not perform to the stated specification and support does not provide warranty coverage to provide drives that do perform according to specification. No change to any RAID controller will make it operate according to specification as the performance issue is specific to the drive and not the RAID controller.

    • fluke
    • 12 years ago

    Maybe this article should be “Seagate sells out” instead of “speaks out.” I would like to thank the author, Cyril Kowaliski, for pointing out again the major issues that Mike Hall has chosen to ignore. The real issue isn’t if Seagate drives sometimes has problems since every hard drive manufacturer has problematic hard drives. Also the biggest issue isn’t if the problem is just to a small group of hard drives. The major issue that Mike Hall fails to answer is at the end of the day does Seagate hard drive actually come with a real warranty or toilet paper that just happens to have “5 years” written on it?

    When a DVD player fails to function to specification, the manufacturer’s support department may ask for side information such as what TV it is connected to or what audio amplifier. However, that information is never used by the manufacturer for ignoring a violation in supported specification.

    Seagate has made it clear that it will use information provided by the customer in good faith to invalidate honoring it’s so-call “5-year” warranty. At the end of the day, Seagate has a 0-year warranty and WD has a 3-year. The Barracuda 7200.11 has a stated random read seek time of “<8.5 msec” and a stated random write seek time of “<10.0 msec” with an average latency of “4.16 msec.” With that type of specification, the drive should never go unresponsive for over 100msecs (10x the stated maximum write seek time)! The fact that Seagate does not want to support GNU/Linux is not “unfortunate” because it is not support of GNU/Linux that is being requested. Instead, it is support of Seagate’s stated specifications for the device that should be provided support. The fact that GNU/Linux is the scape-goat for failing the stated specifications by over ten times also is not “unfortunate” but inexcusable and unacceptable.

    What is even worse is to scape-goat on RAID when the product is advertised for the application of “Desktop RAID.” At first it seems like Mike Hall ignored this issue as well. But it appears that it is worse that that! He states “this does not result in data loss nor does it impact the reliability of the drive.” For Desktop RAID (which in my experience ends up usually meaning RAID1), having both drives listed as unresponsive by the RAID does result in data be left unwritten and data lost. So Mike Hall has chosen to ignore the reality of the situation and spin this as if the problem does not really exist. In my case, the RAID issue only comes up when the drive is unresponsive for over 100ms (again, over 10 times the stated specification). Not only does this cause problems with the RAID array going to degraded and then failed, but has also sometimes required more than one attempt to perform a rebuild of the array.

    Normally I look down on people that turn to lawyers as the answer to working with companies. But in this case, it really seems like Seagate is doing everything it possibly can with it’s effective 0-year warranty support to try to encourage a class action. The fact they are working on a firmware fix is nice but should not replace providing actual warranty coverage. If the fix is not available now, then the drives should be replaced and once the firmware update is available then they can re-issue those drives as refurbished for future warranty issues.

    Based on my previous experience with Seagate, I would have liked to say the official update from Seagate would be that they would honor warranty replacement for any major deviation from specification regardless of OS or use of RAID. It appears instead that Mike Hall is demanding customer just wait an unspecified time period for a firmware fix to be issued. And between the lines of Mike Hall’s response is the message “NEVER trust our so-called ‘warranty’ ever again!”

    Thanks Mike–at least we now know officially where we stand.

      • MadManOriginal
      • 12 years ago

      That was a heck of a long post but what are you talking about with lack of warranty or ‘0 year warranty?’ I don’t read anything that even nearly implies that in the response. Refunds, if the end user chooses, would be through the retailer. Seagate will either provide an updated firmware if that fixes the issue. If a new firmware can’t fix it then another company in the storage subsystem needs to correct somehting, it would be unfortunate if this took a long time and the end user is beyond a refund time period but that’s not a problem with Seagate directly.

        • fluke
        • 12 years ago

        Failure to provide a replacement drive due to use of GNU/Linux or RAID is a failure to honor the warranty. If they are willing to say that a drive that takes over 100ms to respond doesn’t require replacement just because they did not intend customers to use the drive with RAID (even after advertising for it to have use with RAID) then they have decided to terminate the warranty.

        If during the first 3 years of use of a WD hard drive the device grossly deviates from the stated specifications then the 3 year warranty means it will be *REPLACED*.

        The Barracuda 7200.11 has been out less than a year and Seagate is making excuses as to why they are failing to replace their drives that grossly (by over ten fold) deviate from the stated specification. Therefore, the amount of time that Seagate is providing replacement as required by the warranty is *0* years. At the end of the day, the Barracuda 7200.11 does not have a true 5 year warranty in the same sense that a WD product comes with a 3 year warranty. The time period that Seagate performed the replacement of the defective product is less than a year. Therefore it is a 0-year warranty.

        The fact that Seagate has chosen GNU/Linux or RAID as the reason for failing to provide a warranty does not matter any more than if Seagate discovered a cow had jumped over a blue moon. Support for GNU/Linux or RAID is not being requested. Support for Seagate’s own stated specifications is being requested and then rejected.

        While you are correct that it would be up to the retailer and not Seagate to provide a refund or store credit. But the claim of providing a 5-year warranty should imply being able to have a device that works according to specification or to get a replacement during the warranty period with one that does work accordingly.

        The facts are:

        1) GNU/Linux users are being refused warranty service by Seagate and Mike Hall provides no stated correction on the part of Seagate regarding this problem.

        2) RAID users are being refused warranty service by Seagate and Mike Hall provides no stated correction on the part of Seagate regarding this problem.

        3) RAID users are suffering data lost and reliability because of the excessive delay with the Seagate product. Mike Hall has claimed the problem just does not exist.

        This is a very damning situation for Seagate/Maxtor and customers should consider these facts when making future purchases.

      • toyotabedzrock
      • 12 years ago

      I’m sure that the response was given by a low level person, the kind all companies use, who has limited knowledge about computers.

      I have used Seagate drives in my systems and the systems of my customers for 10+ years and they have always proven reliable, i still have a working 9.6GB Cheetah.

      I’m sure they know they fumbled and want to make sure they have a real verified fix before releasing a new firmware. Also consider when asked by Tech Report they admitted it quicker than i would expect of any other company would these days.

      Also I would like to point out that when you receive a hard drive that is either dead or dies very quickly there is a good chance it was mishandled during shipping. Or even dropped off a shelf if you got it from best buy, always grap one near the back of the shelf and look for damage to the box.

      To jafa,
      Are you using an Intel raid controller and the matrix software? I have noticed in a RAID 1 array it is slightly over protective, and even the most insignificant disruption will cause it to start checking the drives, and usually it will find nothing is wrong.

    • pedrofdmp
    • 12 years ago

    I have a 200gb sata Seagate drive that does the exact same thing using the following config:

    XP x64
    C2D + GA-P35-DS3 rev2.0 P35 ICH9 AHCI (hacked to work only ICH9R supported on XP)

      • continuum
      • 12 years ago

      Uh, your drive isn’t one of the models in question… I would take that issue up with Seagate directly rather than here. =P

    • DrDillyBar
    • 12 years ago

    Looking forward to a new firmware. But turning off drive caching seems to have gotten rid of the problem (with a performance hit of course).

    • MadManOriginal
    • 12 years ago

    Hmm to start a fantastic flame war or not? Meh…at least I learned another thing that can be added to the ‘will get long replieson TR front page comments’…mildly insulting Linux zealots.

      • ludi
      • 12 years ago

      I still want to know what girl you were smarting about when you got involved in that exchange.

        • MadManOriginal
        • 12 years ago

        The particular post you’re talking about was the one where I went off about calling CS reps ‘flaks,’ I did ask what that term meant and whether it is derogatory or not but never got an answer. Having worked CS in the past for a company that was quite strict with some policies and also pushed very hard to maxmize ‘production’ (read: push people off the phone as quickly as possible) I know what it’s like on the other end and that some customers call with a bad attitude to start and just make the job a PITA. So it was about customers being mean to CS reps.

      • no51
      • 12 years ago

      Linux fanboys make Macolytes look normal. Case in point, I was lurking around the Blender forums for some tutorials and got into a thread where this one guy was going ape-bananas over Microsoft contacting Blender as part of their open source push or something. You’d think that Microsoft went over to their house and killed everyone they knew or ever had contact with.

    • Dposcorp
    • 12 years ago

    Sometimes the support monkey the gets the call talks out of his rear, but at least the company as a whole appears to try and do whats right.

    Seagate and WD have both been great for me.

    • Snake
    • 12 years ago

    Oooh, big surprise :rolleyes: [/sarcasm]

    After multiple Seagate external drives that have failed on me, and reports from a coworker that his friend has had the same “experience” (multiple drive failures), I recently switched to WD.

    What was happening? The drives seemed to time out on writes, thereby corrupting portions of the FAT. The “Seagate Beep” (when the drive fails the write) should have a Wikipedia entry all it’s own.

    Then again, I’ve had 3 out of 4 Samsung drives fail, so what’s to expect in modern drives anyway…??

      • eloj
      • 12 years ago

      FAT? What is this, 1998?

        • moose17145
        • 12 years ago

        lol i was thinking the same thing. I wonder if a FAT format could even handle this sized drive? I thought for drives this big you needed to use the NTFS format (which uses a MFT over a FAT). I know FAT32 format can handle drives way larger than 32 GB (despite what MS would have you believe), but i didnt think that the standard could handle anything this large.

          • Forge
          • 12 years ago

          “This allows for drive sizes of up to 8 terabytes with 32KB clusters, but the boot sector uses a 32-bit field for the sector count, limiting volume size to 2 TB on a hard disk with 512 byte sectors.”

          So 8TB FAT32 limit, with a 2TB limit for practical usage.

          Now you see why many multi-OS folks got hacked off when MS aborted FAT32 early to push people off 9X (a good thing, but the ends don’t outweigh the means).

          NTFS does outrun FAT32 with technical advantages once you pass a few hundred GB, but it’s viable much farther than that.

          FWIW, NTFS has a bug that’ll start showing up soon as soon as folks start making single partitions >2TB in size.

      • ludi
      • 12 years ago

      So far all of my Seagate drives are fine, although the last new disks I purchased were a pair of SATA 250GB units nearly three years ago.

      I do have three refurbs (explicilty labeled as such) that I got for free as my reward for the trouble of RMA’ing three dead units in a friend’s stash earlier this year, though, so we’ll see what those do. One is a primary disk in my Media PC and the other two are secondary storage drives in other computers.

        • Palek
        • 12 years ago

        I have been buying Seagate drives for the last five years and so far none have failed on me. The first one, a Barracuda IV 80GB, is still going strong after over five years.

          • xastware
          • 12 years ago

          barracuda IV never grove up above 60gb, and in thpse days series 4 & 5 were the most reliable 7200rpm drives so do mess it up with nowadys crrapy drives. since they wider up they business slipstreaming famous maXtor facilities under their umbrella they turn out to produce extremely bad products. every crap has it buyer, sort of seak moto.
          oh i can not to remember how proudly they announcing (in 2004) they have tmr tech that will push mechanic hdds up to 50TB on 5 platters, an look at them now just milking the end-users with their fuckinly unfair firmware policy. most of that crappy hotties in fact dont have mechanical but beta ULTRA-RESTRICTING firmwares so crappy shitgate could proprly segment market for their pleasure.

      • DrDillyBar
      • 12 years ago

      This is the first time I’ve had an issue with Seagate, and if the issue will be fixed soon, I’ve no problem with disabling caching for now. 1.5TB FTW!

    • adisor19
    • 12 years ago

    Seagate needs to get burned for this as hard as possible ! As much as i approve of their 5 year warranty, i totally dislike the way they’re handling this mess. I’ll wait to see how this will pan out and they better do something fast or they’ll lose another customer..

    Adi

      • moose17145
      • 12 years ago

      To take the bait, or not to take the bait… that is the question…..

        • ludi
        • 12 years ago

        I think that’s caviar on toast points. Small in the grand scheme of things, but rather tasty. Choose carefully.

    • jafa
    • 12 years ago

    The official statement is slightly misleading…

    1) When the problem occurs all hard drive operations stop until the OS times out the ATA command – typically 30 seconds. This results in the computer freezing for 30 seconds.

    2) The problem can result in data loss if using a RAID system. Depending on the OS/RAID configuration the problem may cause a RAID system to think the drive has died. The RAID system automatically removes the drive and continues to run degraded (as designed). 20 minutes later when another drive exhibits the problem the RAID system drops the second drive and dies.

    3) The problem may be a systematic problem rather than a small number of drives – all drives have I tested running the SD17 firmware have exhibited the problem.

Pin It on Pinterest

Share This