My server is killing hard drives

[sleonard]

C: is a 250GB WD Black drive never had a problem with it.

D:, E:, and F: were 1TB Seagate drives. D: was for recordings and the few vids and ripped DVD's I had. E: and F: were my music collection.

I replaced D: last spring with this drive. I have never had a problem with it.

Last month I bought two more of the above WD AV drives to replace the two remaining Seagates. I moved the Seagates to another PC and installed the WD's. After formatting them I started to copy my music collection to the new F: drive and found the speed was incredibly slow. After troubleshooting I determined that one of the Seagates was bad.

1) When anything was copied to or from that drive the speed was 20-25 Mb/s at best.
2) CPU use went way up and was found to be HW interupts via Process Explorer.
3) Seagates HD utility told me that the firmware needed updating and research found the model was infamous for shipping with bad FW.
4) The HD also failed the extended tests of that same utility.

While disc access was severly hampered the actual data seemed to be intact so I let the copy crawl along for nearly a week in order to save my data. After I updated the FW on both Seagates and they both now appear to be fine even passing the extended tests.

Both the new WD drives also passed all the tests and appeared to be working. For about a week. Then one of them began showing the same symptoms of the bad Seagate. Slooooow Windows copies to or from the drive. Subsequent testing also showed the high CPU use by HW interupts and it failed the extended tests.

I thought I had just gotten the occasional bad drive an got an RMA from WD. I copied all the data to another drive just in case and waited for the replacement to arrive.

When it did arrive, I installed it, formatted it, ran every test in both the Seagate and WD utilities. It passed and worked perfectly. For about a week.

Now the RMA replacement is showing the same symptoms as is the second drive I purchased from NewEgg.

It now seems as if something in my server is killing any drives connected in the E: and F: positions. (the original c: and d: are still fine. the original d: is the same model as the two new drives that keep dying).

There is only two connections to SATA drives, power and data. So I either have a bad PSU or bad SATA controller on the MB. The Intel motherboard monitor utility reports all power in specs but I'm not sure it monitors every aspect of the power that might cause damage and it is easy to see how a bad PSU could fry a drive.

OTOH, it could be that the bad Seagates fried something on the MB that is now being passed back to any replacement drives. That and in all these cases the data itself appears to be intact. It is just that getting data onto or off of the drive is affected. This could be a SATA controller problem.

I guess it could be that I just got "lucky" and received 3 drives all from the same bad batch but I am not leaning this way. Seems unlikely at best. Especially since two came from NewEgg and one directly from WD as an RMA.

I'm not sure how to proceed. I do know that I don't want to put any more drives into this server until I know for sure what the root cause is. I'll just end up with more dead drives to RMA. How can I figure out for sure what is killing the drives?

Or do I just scrap both the PSU and the motherboard at the same time and take no chances?


Thanks,

S

[GKusnick]

I think it's fairly unlikely that the controller is at fault. My first suspect would be the PSU, followed by some sort of overheating/airflow issue. You can get a PSU tester for about $20 that should tell you if your supply is within spec. You might also look for a cheap digital thermometer with a remote probe that you can tuck into your case to monitor drive temperatures.

[Spectrum]

I agree, this does sound like a power supply issue. The C and D drives may be on a seperate rail on the PSU which is working ok and the rail that E/F are on is malfunctioning. That's just a guess though

On another note, I went through about 6 1TB seagates awhile back before I got two that didn't die. I just replaced the critical drives with WD's and kept RMA'ing the Seagates and and never looked back. The replacements for the replacements for the Seagates are now just used for non critical data storage. Bad batches are popssible but when you cross manufacturer bounderies like that, the odds diminish significantly

[sleonard]

Originally Posted by GKusnick

I think it's fairly unlikely that the controller is at fault. My first suspect would be the PSU, followed by some sort of overheating/airflow issue. You can get a PSU tester for about $20 that should tell you if your supply is within spec. You might also look for a cheap digital thermometer with a remote probe that you can tuck into your case to monitor drive temperatures.

Airflow/Temperature has been found to be OK. Otherwise, the same as I had been thinking.

Originally Posted by Spectrum

I agree, this does sound like a power supply issue. The C and D drives may be on a seperate rail on the PSU which is working ok and the rail that E/F are on is malfunctioning. That's just a guess though

On another note, I went through about 6 1TB seagates awhile back before I got two that didn't die. I just replaced the critical drives with WD's and kept RMA'ing the Seagates and and never looked back. The replacements for the replacements for the Seagates are now just used for non critical data storage. Bad batches are popssible but when you cross manufacturer bounderies like that, the odds diminish significantly

I had the same guess about them being on the same rail. To find out I will have to get a multimeter. You're also right about the cross manufacturer ruling out the bad batch theory. So a bad power supply is looking like a strong possibility. Hopefully after a quality PSU is installed the drives start performing normally instead of sustaining permanent damage.

Yet another idea is that something in XP such as a single partition spanning the entire 2TB capacity is causing this. XP's partition limit is 2 TB I think, and the drives are 2 TB so it could be that the drive is slightly larger than XP's limit and that just breaks something. If that's the case then 2 partitions should solve the problem.

S

[mayamaniac]

What PSU do you have? I'm not sure if its bad or just under powered. Have you had any other system issues beside the failing hard drives?

There's a PSU calculator you can use to see what type of PSU you will need.

[Spectrum]

2GB??? ^G^T
2TB is an artificial limitation due to MBR and if you use dynamic volumes (shudder) or GPT you can create much larger volumes: 2^(32)−1 clusters
Source.
Of course this is for SP2+ SP1 was limited to 128GB

[sleonard]

Originally Posted by Spectrum

2GB??? ^G^T
2TB is an artificial limitation due to MBR and if you use dynamic volumes (shudder) or GPT you can create much larger volumes: 2^(32)−1 clusters
Source.
Of course this is for SP2+ SP1 was limited to 128GB

Oops, sorry. Not paying enough attention. I meant TB not GB. OP corrected.

S

[Savage1701]

You have several possible issues:

Main one: You are using an advanced format drive in this family of drives. Unless you are on Win 7 or Vista, every other OS needs jumpers set to handle the conversion back to 512 byte sectors they know and love. Some insist that jumper still slows things down. I'm going off the link you posted, and I did not see your OS you were using.

These drives are just plain slow and try to use Green tech to lower power consumption, BUT, I have experienced drive slowdowns with improper settings of spread spectrum clocking (should probably be off), using a WD utility to make sure the write-back cache is enabled at the drive BIOS level, making sure that the drive's write-back caching is on at the OS level, and I have also had NCQ turned on and it slowed my 60-80MB/s drive to sub-20MB/s speeds.

These AV drives get poor reviews because they spin slowly and so have inherently lousy transfer rates. Google 'em and you will see what I mean. WD is just pawning off slow drives by giving them an exotic name. Don't drink the Kool-Aid on this one.

If you are using XP, have an Advanced Format variant, and not setting these drives per WD's instructions, you will have a slow drive.

[JetreL]

As long as the system is stable and not jostled very often while running, typically 80-90% of HD failures are caused by heat or bad power. Trying to alleviate this issue by swapping out the power supply would be the first step for me in troubleshooting this issue.

[heffe2001]

You may also want to look at your sata cables.. I've had a few go bad on me over the years, and they cause all kinds of odd problems (diagnostic failure, slow transfers, etc). Swapping out the cables fixed the issues in my case.

[sleonard]

Originally Posted by Savage1701

You have several possible issues:

Main one: You are using an advanced format drive in this family of drives. Unless you are on Win 7 or Vista, every other OS needs jumpers set to handle the conversion back to 512 byte sectors they know and love. Some insist that jumper still slows things down. I'm going off the link you posted, and I did not see your OS you were using.

These drives are just plain slow and try to use Green tech to lower power consumption, BUT, I have experienced drive slowdowns with improper settings of spread spectrum clocking (should probably be off), using a WD utility to make sure the write-back cache is enabled at the drive BIOS level, making sure that the drive's write-back caching is on at the OS level, and I have also had NCQ turned on and it slowed my 60-80MB/s drive to sub-20MB/s speeds.

These AV drives get poor reviews because they spin slowly and so have inherently lousy transfer rates. Google 'em and you will see what I mean. WD is just pawning off slow drives by giving them an exotic name. Don't drink the Kool-Aid on this one.

If you are using XP, have an Advanced Format variant, and not setting these drives per WD's instructions, you will have a slow drive.

I don't have an advanced format model but I am running XP. I have three identical models one of which is several months old and still operating 100%. I did not do anything different when installing, partitioning, or formatting them in XP.

I don't really care about the reviewed speeds being slow compared to other drives. I care that I'm not even getting close to those speeds. The integrity appears to be OK and the data is safe for now but accessing them makes floppy drives look like speed demons. If I could get the reviewed speeds I'd be happy as they're more than fast enough for the job I'm giving them. What I am really looking for is longevity and that ain't working out so far.

Originally Posted by JetreL

As long as the system is stable and not jostled very often while running, typically 80-90% of HD failures are caused by heat or bad power. Trying to alleviate this issue by swapping out the power supply would be the first step for me in troubleshooting this issue.

I bought a new power supply and installed that last weekend. The bad drives are still bad but that is not unexpected. I don't have a way to test the former PS so I don't really know if the problem is fixed for sure or not. I'm still not to keen on testing with some more brand new drives. If it turns out that the SATA controller on the MB caused the problem then I will have sacrificed another drive.

Originally Posted by heffe2001

You may also want to look at your sata cables.. I've had a few go bad on me over the years, and they cause all kinds of odd problems (diagnostic failure, slow transfers, etc). Swapping out the cables fixed the issues in my case.

I do have some more cables and will try that tomorrow but I don't have much hope for this since all the bad drives worked perfect for at least a week before showing any symptoms.

S

[drewg]

I wonder if something is making your computer downshift to PIO mode for your "slow" drives. There was recently a posting about this in the beta forum:

http://forums.sagetv.com/forums/show...54&postcount=1

Drew

[davephan]

I have 9 Western Digital 2 TB drives, and had some trouble with the drives, but I don't consider the drives to be a bad design to be avoided. The drives are Western Digital 2 TB green WD20EARS and WD20EADS, probably similar to your WD 2 TB green drives.

Two drives failed when connected to a bad power supply. It turned out the one of the connectors on the power supply was supplying bad voltage, I assume too high. I did not measure the voltage. Both drives immediately sparked and had a burnt smell after that. Both drives were immediately dead. That was a bad power supply problem. I replaced both drives that were destroyed by the bad power supply and I also replaced the bad power supply.

I also had one drive that ran for a day and went bad. I returned the drive to the store and the replacement has been working fine since.

My 9 drives have been all been running now for 6 months to over a year without problems. The drive temperatures range from 25c to 30c for each drive mounted in Thermaltake iCage A2309 drive cooling cages. The product used to be available from Newegg, but is no longer available. It is still available from Amazon. The density is much lower than the 5 in 3 cages, so the airflow should be better, and it probably keeps the drives cooler than the high density cages. I have one drive that is running without a cooler cage at 44c. I tried adding a fan near the hot drive, but it did not lower the temperature at all. Maybe the drive cage will help cool down that drive or maybe that drive has a problem and might need to be replaced.

Here is a link to the Thermaltake iCage A2309 drive cooling cages:
http://www.amazon.com/gp/product/B00...ef=oss_product

What are your drive temperatures?


Dave

[dcardellini]

Originally Posted by davephan

.....My 9 drives have been all been running now for 6 months to over a year without problems.

Dave

Not to take this thread off topic.... but on another thread close by......I ran 11 WD20EADS's for a year problem-free, and suddenly in past three weeks lost three.

As my drive temps run 37-41 degC,, I am wondering if temperature is a big factor here. I did read a great article not too long ago about drive temperatures...and it was noted that running cooler does not necessarily increase life....in fact.....just the opposite. The article indicated that there was an optimal temp range (I wish I could find the link).

Are you running your drives 24/7?

[sleonard]

Originally Posted by drewg

I wonder if something is making your computer downshift to PIO mode for your "slow" drives. There was recently a posting about this in the beta forum:

http://forums.sagetv.com/forums/show...54&postcount=1

Drew

Both of the failing drives were in PIO mode. I had concluded that this was so because the drives had problems. As a result of your post I went back and looked at this again thinking that maybe something else caused Windows to put them in PIO mode and that, in turn, caused the poor performance and failed tests.

BIOS is the place so I booted into setup and looked at the controller/disc setup. What I found was that all the options were set correctly but just for the sake of it I changed the controller from IDE to AHCI. As I expected the PC would no longer boot at all so I ended up changing it back to IDE.

Lo and behold, when I got back into Windows both discs were in Ultra DMA mode 5. I started the WD utility extended test on one of them and it is running. (before it would struggle for about 5 minutes while the time remaining clock gradually increased to dozens of hours. At about five minutes it jumped to 3500 hours remaining and I would end up canceling the test).

I don't know what it was that was the root cause but toggling that BIOS setting seems to have unstuck a bit somewhere and they are now running OK. Hopefully it stays that way.

Thanks,

S

[davephan]

Originally Posted by dcardellini

Are you running your drives 24/7?

Yes, I am running the drives 24 x 7. Some WD 2 TB green drives are in the unRAID server, some are still stand-alone. I have the unRAID server set to not spin down the drives.

Dave