locked
What Hard Drives are the Best? RRS feed

  • Question

  • I am tired of failing hard drives.  I have 9 in my WHS and I have replaced three of them so far this year.  Now a 7 nmonth old 1TG drive has failed. All of the hardware is less than one year old.  The real culprits are WD drives.  Anybody have opinions on the best ones for replacements?  Sorry if this is in the forum somewhere.  To p)&)*&*)&*&(*() to look!
    Friday, May 15, 2009 6:22 PM

Answers

  • When I see multiple drive failures over a short period of time, I generally look to other hardware. While every drive manufacturer has a certain percentage of drive failures, those percentages are really fairly similar across the industry. So clusters of drive failures are likely to indicate an environmental issue of some sort, rather than a manufacturer that makes bad drives. 

    Which Western Digital drives have you been buying? What sort of failures have you been experiencing? Is it possible that your power supply is weak? How about cables in your server? There are a lot of possible factors that could affect the drives in your server...

    I'm not on the WHS team, I just post a lot. :)
    Friday, May 15, 2009 6:33 PM
    Moderator
  • I also have a 9-disk, 7TB WHS along with 5 PC's who's disks get a thorough thrashing by my twins who are both studying media and have been doing a lot of video editing over the past year.

    I've got Samsung, Hitachi, Maxtor, Seagate and WD disks and so far, we've had 3 Maxtor (one 320GB and two 500gb - one of which was in an external USB enclosure) and one WD (1tb) disk fail but never had any problems with the Samsung, Hitachi or Seagate disks. Since my experience of the Maxtor disks has been to see them fail in a variety of implementations, I now avoid that brand like the plague. As they say, your mileage may vary and since there's a variety of brands out there who all seem to be able to stay in business, I don't think there's a clear "best" drive to be had.

    More generally speaking, my view is that you're better off trying to minimise known fault-inducing circumstances. Buying RAID-rated disks for your WHS and leaving them on 24/7 is a good first step. A repeated on/off cycle will introduce both thermal and electrical stress on the drives as the drive motor draws maximum current (amps) on start-up and the metal in the drive chassis along with all on-board components will expand/contract - often unevenly - with each heating up/cooling down cycle from being turned on/off. This is actually what kills most consumer electronics over time - electronics and magnetic media prefer a constant temperature.

    That brings us to the next point: cooling - there's a clear correlation between temperature and MTBF (google it!) so adequate cooling for all disks is a cheap way of reducing the risk of failure. A couple of 120mm fans at the front of the enclosure, blowing across the drives and another couple of 120mm fans at the back, evacuating hot air from the enclosure ensures a constant air flow through the box and will also help keep the RAM, CPU, etc. at a constant temperature.

    Tentacle Software's excellent "Disk Management" plug-in (http://www.tentaclesoftware.com/whsdiskmanagement/) can display the internal temperature of each individual drive using the built-in temperature sensor (most SATA and more recent IDE drives have one, although some IDE controller cards won't report this information over the bus). This is an excellent tool in identifying if a drive runs too hot and needs a further fan pointing at it - especially if you go through the trouble of creating a wireframe image showing the physical location of each of your drives. With fans costing a couple of pounds (I get mine from eBuyer, the budget range is fine and cost about a fiver for a 120mm fan and a couple of quid for smaller versions that can be creatively installed where needed using cable ties. Get a few extra for spares, you never know...)

    When it comes to speed, as been said before in this forum, the drive speed is very unlikely to be the bottleneck, even for the slowest drives as all the data has to make its way through a number of components before being sent down the network by the network controller. However, I just put a new Seagate 1.5TB drive in my WHS and the specs for that seems to indicate that it's at least as fast as some of the 10,000 rpm drives but only draws a peak 15W and 6W idle - very useful in a WHS.

    Finally it's probably worth mentioning power provision - especially if you are re-purposing an old computer.

    Make sure the power supply has the oompf to provide the combined peak power needed on boot but does not run at less than 60% capacity when the system idles (most, except for the very expensive, power supplies' efficiency drops dramatically when they're operating at a low percentage of capacity, so you're effectively paying for electricity to be converted into heat in the power supply. Most modern drives max out below 20W and idle around 5-10W but be mindful of where this wattage is created: google your drive to find how many amps (A) of 5v and 12v the drive needs and calculate the wattage per voltage rail in the power supply. ( amps x volts = watts, so for example 1.25A at 12v is 15W while 1.25A at 5v is 6.25W) The power supply should have printed on it a) the maximum wattage, i.e. 550W but will often also indicate max Amps or Watts per voltage type and whether there are several rails (i.e. sources) of any voltage, usually 5V. If your power supply has several 5V rails, make sure your drives are equally divided (by wattage or amps) over the different rails as an idling rail will waste electricity to make heat while the used rail might be overloaded and unable to supply the required amps, leading to mysterious and apparently random faults/errors.



    Sunday, May 17, 2009 9:54 AM

All replies

  • When I see multiple drive failures over a short period of time, I generally look to other hardware. While every drive manufacturer has a certain percentage of drive failures, those percentages are really fairly similar across the industry. So clusters of drive failures are likely to indicate an environmental issue of some sort, rather than a manufacturer that makes bad drives. 

    Which Western Digital drives have you been buying? What sort of failures have you been experiencing? Is it possible that your power supply is weak? How about cables in your server? There are a lot of possible factors that could affect the drives in your server...

    I'm not on the WHS team, I just post a lot. :)
    Friday, May 15, 2009 6:33 PM
    Moderator
  • All of the drives in the WHS are 500gb WD, less than one year old.  The PSU and other hardware is also about the same age. The PSU is 700 watt,  I am fairly secure in the hardware being in good shape.  I run about 8 desktops and two laptops here and functionally, the WHS has worked fine.  All of the other desktops are running similar hard drive sizes with a couple of 1TB here and there, mainly for media storage.  The one that failed was a Not Added drive in the WHS for storage of recorded TV only, about 900gb worth.  The drive started not responding periodically and finally it stopped all together.  I tried it in two other desktops and it had simply died.
    Friday, May 15, 2009 7:18 PM
  • Just read this comment on Newegg:

    Pros: Huge capacity.
    Cons: Poor quality control. We have ordered 5 of these drives for clients - we are working on RMA #3. The first drive was DOA - click of death. The replacement was fine. Then we ordered two more a few months later - one seemed fine, the other was DOA - once again the COD. A week after installing the seemingly fine drive we are seeing signs of failure (had to rebuild a RAID) and hearing ominous sounds coming from the drive.
    Other Thoughts: I run a computer repair shop. Western Digital has long been a trusted name in hard drive storage - but I wouldn't trust a drive larger than 500GB from them until they get these production kinks worked out. As of this writing the WD10EADS does not have a firmware fix available.... not that it would fix a click of death failure.

    Guess that is the problem with the 1TB WD drives!
    Friday, May 15, 2009 8:05 PM
  • I'm using the Samsung Spinpoint F1 - 750 GB for at least a year now. A few months with the 1 TB drives. Awesome drives at a great price too.

    I'm looking to go to 1.5 TB, but the reviews on the Seagate (even today with the firmwire fixes) aren't good. Lots of DOA with those and the WD. Sadly I'll stick to the Samsung 1 TB for now
    Friday, May 15, 2009 9:23 PM
  • I also have a 9-disk, 7TB WHS along with 5 PC's who's disks get a thorough thrashing by my twins who are both studying media and have been doing a lot of video editing over the past year.

    I've got Samsung, Hitachi, Maxtor, Seagate and WD disks and so far, we've had 3 Maxtor (one 320GB and two 500gb - one of which was in an external USB enclosure) and one WD (1tb) disk fail but never had any problems with the Samsung, Hitachi or Seagate disks. Since my experience of the Maxtor disks has been to see them fail in a variety of implementations, I now avoid that brand like the plague. As they say, your mileage may vary and since there's a variety of brands out there who all seem to be able to stay in business, I don't think there's a clear "best" drive to be had.

    More generally speaking, my view is that you're better off trying to minimise known fault-inducing circumstances. Buying RAID-rated disks for your WHS and leaving them on 24/7 is a good first step. A repeated on/off cycle will introduce both thermal and electrical stress on the drives as the drive motor draws maximum current (amps) on start-up and the metal in the drive chassis along with all on-board components will expand/contract - often unevenly - with each heating up/cooling down cycle from being turned on/off. This is actually what kills most consumer electronics over time - electronics and magnetic media prefer a constant temperature.

    That brings us to the next point: cooling - there's a clear correlation between temperature and MTBF (google it!) so adequate cooling for all disks is a cheap way of reducing the risk of failure. A couple of 120mm fans at the front of the enclosure, blowing across the drives and another couple of 120mm fans at the back, evacuating hot air from the enclosure ensures a constant air flow through the box and will also help keep the RAM, CPU, etc. at a constant temperature.

    Tentacle Software's excellent "Disk Management" plug-in (http://www.tentaclesoftware.com/whsdiskmanagement/) can display the internal temperature of each individual drive using the built-in temperature sensor (most SATA and more recent IDE drives have one, although some IDE controller cards won't report this information over the bus). This is an excellent tool in identifying if a drive runs too hot and needs a further fan pointing at it - especially if you go through the trouble of creating a wireframe image showing the physical location of each of your drives. With fans costing a couple of pounds (I get mine from eBuyer, the budget range is fine and cost about a fiver for a 120mm fan and a couple of quid for smaller versions that can be creatively installed where needed using cable ties. Get a few extra for spares, you never know...)

    When it comes to speed, as been said before in this forum, the drive speed is very unlikely to be the bottleneck, even for the slowest drives as all the data has to make its way through a number of components before being sent down the network by the network controller. However, I just put a new Seagate 1.5TB drive in my WHS and the specs for that seems to indicate that it's at least as fast as some of the 10,000 rpm drives but only draws a peak 15W and 6W idle - very useful in a WHS.

    Finally it's probably worth mentioning power provision - especially if you are re-purposing an old computer.

    Make sure the power supply has the oompf to provide the combined peak power needed on boot but does not run at less than 60% capacity when the system idles (most, except for the very expensive, power supplies' efficiency drops dramatically when they're operating at a low percentage of capacity, so you're effectively paying for electricity to be converted into heat in the power supply. Most modern drives max out below 20W and idle around 5-10W but be mindful of where this wattage is created: google your drive to find how many amps (A) of 5v and 12v the drive needs and calculate the wattage per voltage rail in the power supply. ( amps x volts = watts, so for example 1.25A at 12v is 15W while 1.25A at 5v is 6.25W) The power supply should have printed on it a) the maximum wattage, i.e. 550W but will often also indicate max Amps or Watts per voltage type and whether there are several rails (i.e. sources) of any voltage, usually 5V. If your power supply has several 5V rails, make sure your drives are equally divided (by wattage or amps) over the different rails as an idling rail will waste electricity to make heat while the used rail might be overloaded and unable to supply the required amps, leading to mysterious and apparently random faults/errors.



    Sunday, May 17, 2009 9:54 AM
  • Thanks for all the help.  I am going to try a Samsung Spinpoint F1 - 1TB for the first time.  I am going to place it in an external enclosure for Recorded TV only and see what happens.  Thanks again. 
    Tuesday, May 19, 2009 7:41 PM