Results 1 to 10 of 10

Thread: Enterprise class hard drives

  1. #1
    Join Date
    Jul 2003
    Location
    Sydney, Australia
    Posts
    5,644

    Enterprise class hard drives

    Hey guys I need some help. I've been trying to find some extremely reliable SATA internal hard drives. They seem to be known as Enterprise class or RAID-class HDDs.

    I have found the Western Digital WD2002FYPS and the Seagate Constellation ST32000644NS. However, I have had several Seagate drives go bad on me so I have been buying Samsung 1TB SATA drives for builds in the last year. These are not Enterprise class but are still 7200rpm and 32MB cache internal drives. In the last week two of these Samsung HD103UJ 1TB drives have failed in our Server PC. They were part of a RAID 5 array totalling 8TB. The Server is running an Areca ARC-1222 8 port low profile PCI-E x8 SAS RAID controller that set me back nearly AUD800.


    What I need is extreme reliability.


    Any help appreciated.


  2. #2
    Join Date
    Sep 2010
    Location
    Leiden, the Netherlands
    Posts
    4,384
    Hi Vaughan,

    Can you change your RAID-5 array into a RAID-6 one? This article http://storageadvisors.adaptec.com/2...ions-for-raid/ claims a *much* longer Mean Time To Data Loss (MTTDL) for that kind of setup.


  3. #3
    Join Date
    Jul 2003
    Location
    Sydney, Australia
    Posts
    5,644
    Thanks for the link Dirk. Time to investigate this.


  4. #4
    NeoGen's Avatar
    NeoGen is offline AMD Users Alchemist Moderator
    Site Admin
    Join Date
    Oct 2003
    Location
    North Little Rock, AR (USA)
    Posts
    8,451
    As far as I looked up it seemed like RAID-6 is just one step further ahead of RAID-5, in that it stripes data and parity information between drives in the same fashion, but it will add an extra parity strip so you will need at least 4 drives to build a RAID-6 setup instead of 3 drives for a RAID-5.
    I didn't see anything about higher hard drive reliability but I did see that a RAID-6 system will keep on going and not lose data even when 2 hard drives fail simoultaneously, which RAID-5 can't handle. RAID-5 can support the loss of 1 hard drive, but not 2.

    RAID-6 seems to be rare too, I've searched around for motherboards or PCI RAID cards that support it and there are not that many. The ones that do might cost you an arm and a leg.

  5. #5
    NeoGen's Avatar
    NeoGen is offline AMD Users Alchemist Moderator
    Site Admin
    Join Date
    Oct 2003
    Location
    North Little Rock, AR (USA)
    Posts
    8,451
    I did a quick search on google for enterprise class hard drives and came up with an article from 2007 on Tom's Hardware website:
    http://www.tomshardware.com/reviews/...ives,1575.html

    By the examples they show, it seems that the "Enterprise Class" hard drives are the ones that spin at 10K and 15K RPMs, and have very low capacities (up to 300Gb).
    Professional drives utilize Serial Attached SCSI or Ultra320 SCSI interfaces, they spin at 10,000 RPM or 15,000 RPM and they are highly optimized for performance. High-capacity storage solutions are built on enterprise-ready Serial ATA hard drives, such as the Seagate Barracuda ES.

  6. #6
    NeoGen's Avatar
    NeoGen is offline AMD Users Alchemist Moderator
    Site Admin
    Join Date
    Oct 2003
    Location
    North Little Rock, AR (USA)
    Posts
    8,451
    And one more comment if I may... when acquiring hard drives, don't get them by the mail or UPS or similar delivery systems. Hard drives are sensitive to motion and unless they are well protected and carefully handled during transportation (which 9 out of 10 times I bet they are not) they will come with problems.

    I know this because in my previous job I remember I had to build some storage servers and they ordered half a dozen drives that were delivered to us by UPS. Two of them had clicking sounds as soon as they were plugged in and a third one seemed to be doing fine until a full scan was performed and it started showing bad sectors.

    So... I don't know how things work out there, but every since then I always get my hard drives from local stores. It seems they take much less of a beating when they are shipped from the manufacturer to the stores, than from the stores to the homes... I can only suppose that the shipping from manufacturer to stores is done under appropriate conditions, while stores don't have as much care.

  7. #7
    Join Date
    Mar 2007
    Posts
    638
    I have had less failures with systems that have an intake fan blowing across the hard drive. I think this works better then the water blocks on hard drives as they are usually after the water temp is hotter then ambient air. I have been going between WD Green power hard drives and Seagate hard drives depending on what is on sale. They are mostly in the 500GB range so far. The western digital black series is something you may want to look into as they had a 5 year warranty if i remember right. Best of luck on finding a good setup as DATA loss can be a huge pain.



  8. #8
    Join Date
    Jul 2003
    Location
    Sydney, Australia
    Posts
    5,644
    Ended up removing the dead drive from the RAID 5 array.

    Now working on a backup to the server using an ultra micro VIA box that my son imported from Canada without even telling me!!! I just got told hey Dad I need two 2TB WD caviar black SATA 2 HDDs and 2x 2GB Kingston DDR2-800 so-dimm RAM sticks for the new backup box. I must admit its a cute little computer, about the size of a loaf of bread. Its going to run Ubuntu.


  9. #9
    AMDave's Avatar
    AMDave is offline Seeker of the exit clause Moderator
    Site Admin
    Join Date
    Jun 2004
    Location
    Deep in a while loop
    Posts
    9,612
    consumer / non-enterprise HDDs only average about 1 year in the life of a raid pack IIRC.
    there are various differences in the materials and methods used in manufacture that results in a big difference in quality.
    However, if you build that into your plan for the life-cycle of your RAID array and aim to swap HDDs out before they fail, then you can save yourself some trouble AND migrate them out to your boxen as you replace them with new ones.
    That way you get rid of your oldest smallest disks among the boxen first and if you have multiple failures in your raid packs, you can temporarily decom one of the boxen and grab a recently migrated HDD back into the array to keep it spinning while the replacement order is being processed.

    I am using Hitachi HDS72101 SATA HDD's in pairs at the moment.
    Planning on a 6 month replacement cycle.

    8TB huh?
    You should put together some maintenance plans for that kind of infrastructure and maybe part-time hire Doomeva to keep the maint-cycle target dates within acceptable limits.

    Hmm.
    Are we starting to sound like data-center support?

    Can I sell you some ITIL facilitation software?
    And maybe some on-site training for your 'staff'?
    You'll get a free mouse-mat to show off to your friends.
    HA HA HA j/k
    . . . . . ___
    . . . . . . .\___/\______
    . . . . . . . \__AMD___\\__
    ---------------------------------------------

  10. #10
    Join Date
    Jul 2003
    Location
    Sydney, Australia
    Posts
    5,644
    Good idea, I'd like to make sure we have safe data storage so a planned retirement of older drives to the DC crunchers would leave the main server with fresh and large capacity drive. Seeing as the 2TB WD black caviars were not outrageously priced we could end up with 16 TB storage in the server box. Bit of a waste having 1TB or larger drives in a cruncher but the 1TB HDDs are sub $75 now so it doesn't matter much.


Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •