Barracuda XT, Barracuda, Barracuda Green, DiamondMax and other desktop SATA and IDE drives

[mooner]

Very sad day... I've already gotten rid of all my 7200.11 1.5 TB drives... But now my 2TB Barracuda LP drives are now showing intermittent problems. I have 4 of these drives in a gen 2 Drobo. I have 2 drives that are intermittently dropping out. Can't do any real troubleshooting with the drives installed in the Drobo, so all I can do is cross fingers and power cycle the thing. Since the thing is a "black box" (literally and figuratively), I don't even have any ability to see what the drives are (not) doing. Can't tell if they aren't spinning or if they are reporting 0GB or if they're even responding to SATA...

Reboot seems to usually fix, but this just is a very bad situation...

**sigh**

[ByrneIT]

I've found it to be hit or miss on getting 7200.11 drives replaced with 7200.12. If the rep I'm talking to won't do it, I just hang up and call back. I'm now down to only one 7200.11 in use in RAID, after the other 15 failed, including some repeatedly.

One thing I disagree with is that RAID 5 is not "Desktop RAID". To me, the difference between server and desktop is not the RAID level, but the utilization. Just about every high end WS motherboard now supports RAID 5, and this is a very common config for eSata home media sharing.

Instead of using indeterminate marketing speak like "Desktop", if the determining constraint is RAID level, then they should SAY that.

Regardless, in my experience, 7200.11 drives are unreliable as single spindle, RAID 0, RAID 1, RAID 10, and RAID 5. Seagate should recall them, as each one of them is a ticking timebomb waiting to destroy your data.

Message Edited by ByrneIT on 02-22-2010 12:38 PM

[josh1980]

I have a Highpoint RocketRAID 3560 and I am exhibiting the same issues as everyone else.  I am running 16x 1.5TB ST31500341AS with CC1H firmware on a RAID-6 array.  The hard drives show SMART is good and some bad sectors on each drive, but nothing too excessive.  To me, the array will suddenly freeze.  This typically occurs while reading or writing a large quantity of data (don't even think about doing a defrag).  In any case, any attempt to access any drive letters from the array will simply freeze indefinitely.  No hard drives will click, no hard drive will disconnect from the array, no other way to identify any issues with any drives.  The RAID controller will begin returning error messages every 30 seconds to the windows logs stating that the controller initiated a reset. 

You cannot log into the hard drive controller from the computer using the web browser (the normal way to access the controller aside from the controller's BIOS).  I found a VERY interesting fact today.  Since my controller has a built in network card for accessing the card across the network, I decided to access the card remotely.  From another computer I logged into my controller and requested a rescan of all ports.  Guess what happened?  The array immediately began operating just as if nothing happened.  In this particular case the computer froze while copying about 100GB of data from one drive  letter to another (both on the same array).  The copying immediately resumed and continued as if nothing happened.  After copying completed I did an MD5 checksum of all of the copied files, and all files did copy correctly.

It would appear that somehow the hard drive and raid controller stop talking (no clue which component is the actual cause), but by performing a raid controller rescan of all devices this somehow unlocked the issue.  I would like to note that the controller did not see a hard drive disconnect and reconnect during this.

Up until this time, I have used different raid controllers by RMAing the first card, trying a different model of controller, different firmware versions for the controllers, even changing out the SATA cable.  The only way I have been able to unlock the array until now was shutting off the computer using the power switch.  The OS would freeze if trying to shut it down correctly.

Hopefully this will help someone. 

[fzabkar]

josh1980, it does appear that you've found the magic bullet. However I would try to repeat the procedure, just to be certain that the array didn't coincidentally wake up on its own.

[axelkloth]

Josh1980: congratulations on that find. It seems to solve at least the HighPoint RAID controller issues. I am not sure if a blanket statement can be made, though. I have tried an AMCC/3Ware 9690, an Adaptec ICP 5085 and the Intel ICH10-R HostRAID controllers, both with Intel and Adaptec drivers. All of them had issues with the 7200.11 and to some degree with the 7200.12, even the ES.2. They had no issues whatsoever with the Cheetahs, Hitachi GST DeskStars or UltraStars. While I won't rule out firmware problems in a RAID controller, I really have a hard time believing that the 7200.11 disks are NOT at fault. The likelihood that two different dedicated RAID controllers and one HostRAID with two different drivers can have identical problems on one set of disks and show no problems whatsoever with other disks is just to olow for me to believe it.

In either case, it is ridiculous that Seagate's customers conduct a system test without being compensated for it...

But again, kudos to you for discovering your issue.

[josh1980]

@fzabkar - Your logic is correct that I do need to remove the chance that the array coincidentally woke up on it's own.  However, every time I have had the array freeze which always, results in a windows log entry every 30 seconds,  it has always been followed by the log entries signifying that the computer was improperly shut down.  I have gotten my RAID array to start working twice more since my last post.  I am convinced that I have solved the problem with my array, however having to remotely connect to my RAID controller is unacceptable.  If only the firmware could be set up to "auto detect" every 30 seconds or 1 minute.  Then, in the worst case scenario I would only have to wait 30 seconds.

In all honesty, it is possible that there is a "bug" with both the Highpoint 3xxx and 4xxx series controllers, but since they are a different family I tend to exclude that possibility.  I also have a Corsair 1000w power supply, and it is typically loaded to about 250-300 watts when idle. So I definitely am not experiencing an issue with insufficient power.

I had the exact same effect with the following controllers:

RocketRAID 4320 with latest 2 firmwares

RocketRAID 4322 with latest 2 firmwares

RocketRAID 3560 with latest 2 firmwares

Without more information I believe that the issue is with communications between the RAID controller and the hard drive.  Upon resetting the communications (via detecting devices) the issue resolves itself.  I'm sure that the SATA standard probably controls the communications, and because the common denominator seems to be using Seagate hard drives, I tend to believe that Seagate hard drives are not operating per the SATA communication standard(by design or by bug I cannot tell).  I am thinking by design considering how many issues people have had with no responses from Seagate.

I have also considered the possibility that random background EMF is to blame for these situations.  My server case does have plexiglass windows, so it does not form a good faraday cage.  I am considering making my case more protective of EMF to see if this affects my server. If I decide to go through with it I will post my results. 

For those of you unfamiliar with what a faraday cage is, search wikipedia for it.  If you understand what it is explaining you should also realize that having your computer anywhere except inside a case that is not completely shielded within a case made of conductive material and closed is more susceptible to EMF and therefore reliability.  How much does this affect reliability is up for debate.  I consciously chose to have the "cool" plexiglass window because I have had several cases with windows and not had a reliability issue.  Of course, i have never had so many hard drives inside a single case before, and that could make a difference.

Another thought... since it appears to be an issue between Seagate hard drives and only certain controllers, do all of those controllers support hot swap of hard drives?  I wonder if that is related somehow.

Message Edited by josh1980 on 02-26-2010 09:26 PM

[wuZheng]

Hi,

Drives: 2x ST3500320AS; Firmware: SD1A (on both); RAID 0

Controller: ICH10R

Mobo: Asrock P45R2000-WiFi (Intel P45/ICH10R board)

OS: Win7x64 Pro

First of all, although I'm 99% sure it is the drives that are failing, perhaps you guys can decide that for me.

I have the drives in RAID 0 primarily for the performance benefit, as usually is the case with people who use RAID 0. Data redundancy is not of any particular importance to me. Heres the problem though, Windows just hangs, and it always hangs during times of high I/O activity from the array. And I don't mean like hang as in "not responsive windows" but like the OS doesn't even bluescreen, or blackscreen, the screen remains as it was when the error occurred, USB devices suddenly lose all power, but the computer is still on and there are blips of HD activity. 

Intel Matrix Storage manager says nothing is wrong, SeaTools can't read the individual drives, HD Tune can't read the individual SMART values of each drive. Also, defragging has now become a very bad idea, since doing so will hang the machine in the aforementioned way and also potentially do more damage to the array's stability (or lack thereof). As I said, all thing point to the drives as the failure point, but if I'm wrong, please tell me. 

When the machine hangs, a system power cycle always brings the machine back and I can use the machine again.

I'm planning on dropping by a store and picking up a pair of .12s, the reliability (as far I as have experienced) is much much better.

[Borgster]

What a surprise.

[tarknz]

It is very hard for me to get enthusiastic about Seagate - a company whose drives I used to value very highly. I have a relatively new PC (Janaury 2009), set up initally with a RAID 1 configuration of two ST31000340 AS (7200.11 1TB) drives. I have had three 340's fail on me - fortunately replaced under warranty. The last two failures were replaced with  ST31000528 AS (7200.12 1TB) drives. A 528 failed two weeks ago and another this morning - thats five Seagate drives failing in 16 months!!

My first warning in each case was the Intel Matrix Console advising that a RAID drive was degraded.

I am now awaiting for the PC company who built my PC to courier a replacement drive and I have to say - unfortunately - it will be another Seagate because of the warranty. Quite honestly I would rather take my chances with Samsung or WD. When the next drive fails then I will cut my losses and ignore the warranty and replace both with Samsung or WD. I run a small business and cannot afford the risk of losing both RAID drives and my current data. I do a full system back up once a week but that doesn't cover a weeks data should I lose both drives.

My PC has 4 x 120mm fans and is well ventilated.

Come on Seagate - when can I fit Seagate drives with confidence again?

[HughR]

tarknz:

This thread is about physical drive failures, in particular ones associated with loudish noises.  It isn't clear that that is your problem.

Perhaps you should look at this horrible thread http://forums.seagate.com/t5/ATA-and-Serial-ATA-drives/RAID-problems-with-7200-11-drives-unified-thr...

It doesn't really cover 7200,12 drives but it may give you some feel for the scope and confusion about problems with 7200.11 drives and RAID.

Summary: most drive manufacturers seem to want you to buy expensive drives to do RAID.  But even the expensive Seagates seem to run into problems.  Seagate has not really engaged us about what is really going on.