Synology Disk Failure

Discussion in 'Mac Accessories' started by bgd, Aug 19, 2015.

  1. bgd macrumors regular

    Joined:
    Aug 30, 2005
    Location:
    SG
    #1
    I recently had a disk failure on my Synology and was pretty happy with how it was handled.

    I have a DS213+ with 2 x 4TB drives in a raid 0 configuration. I use the NAS for data and TimeMachine backups. Disk 2 failed.

    Reassuringly, with raid 0 and one disk failed, the data was still available until I replaced the failed drive. If needed, I could have copied the data to another drive before replacement. Raid 0 is a little more robust than I thought.
     
  2. MagicBoy macrumors 68040

    MagicBoy

    Joined:
    May 28, 2006
    Location:
    Manchester, UK
    #2
    You sure?

    RAID 0 is striped across both disks. You should have lost all your data.
     
  3. bgd thread starter macrumors regular

    Joined:
    Aug 30, 2005
    Location:
    SG
    #3
    That is what I expected too. Not what happened though.

    Disk had a bad sector and I was able to repair and format using Disk Utility. Perhaps a complete failure might have had a different result.
     
  4. pastrychef macrumors 601

    pastrychef

    Joined:
    Sep 15, 2006
    Location:
    New York City, NY
    #4
    This doesn't make any sense.

    1. One out of two drives died in a Raid 0 array and you replaced the drive.
    2. You repaired and formatted the array using Disk Utility.

    Either of these actions should have resulted in loss of all data on the array.
     
  5. petvas macrumors 601

    petvas

    Joined:
    Jul 20, 2006
    Location:
    Mannheim, Germany
    #5
    Exactly. There is no way that a RAID-0 stripe set can recover from a disk failure.
     
  6. bgd thread starter macrumors regular

    Joined:
    Aug 30, 2005
    Location:
    SG
    #6
    The data was lost once I removed the failed disk. Up to that point the data was available.

    Under raid 0 I would have expected loss of all data on failure, that didn't happen.
     
  7. petvas macrumors 601

    petvas

    Joined:
    Jul 20, 2006
    Location:
    Mannheim, Germany
    #7
    If there was only a bad sector on the disk, then you might have had luck. Data was probably moved to a healthy sector before failure..
    When you removed the second disk, then the stripe set lost all of its data.
     
  8. bgd thread starter macrumors regular

    Joined:
    Aug 30, 2005
    Location:
    SG
    #8
    That's what I was wondering, was I lucky or does the NAS have the ability to move to a healthy sector?
     
  9. pastrychef macrumors 601

    pastrychef

    Joined:
    Sep 15, 2006
    Location:
    New York City, NY
    #9
    You were lucky.
     
  10. MagicBoy macrumors 68040

    MagicBoy

    Joined:
    May 28, 2006
    Location:
    Manchester, UK
    #10
    Yeah, you were lucky.

    I got a couple of bad sectors on a Netgear NAS (ReadyNAS Duo I think?) last year and the RAID 0 dropped out. After a restart and integrity check 99.9% of the files were OK.

    Unfortunately not OK enough to pass the Time Machine integrity check, so it nuked the entire backup. Before anyone goes off on one about backup strategies - it's fine as the computers also backup to a USB drive, hence using RAID 0 on the NAS instead of RAID 1. ;)
     
  11. bgd thread starter macrumors regular

    Joined:
    Aug 30, 2005
    Location:
    SG
    #11
    I will assume I was lucky then. Pity, was hoping the Synology had some inbuilt redundancy to handle bad sectors.

    My reading around the subject has shown that RAID is about availability rather than backup. I'm comfortable having my main back up on RAID 0. I also have a second backup and offsite storage, which is why I didn't need to offload the data before rebuilding the volume.
     
  12. grunty macrumors newbie

    Joined:
    Aug 29, 2014
  13. pastrychef macrumors 601

    pastrychef

    Joined:
    Sep 15, 2006
    Location:
    New York City, NY
    #13
  14. DanClark macrumors newbie

    Joined:
    Dec 25, 2008
    #14
    Let's clear up some misconceptions about RAID...

    First "RAID" is NOT about availability. There are multiple types of RAID. Only some RAIDs are about availability. Each RAID type has it's own purpose.

    Common RAID types are RAID 0 and RAID 1. RAID 0 give you throughput because it is striped. Raid 1 gives you security because it is mirrored - if one drive goes out, the data is still available on the other drive. RAID 0 is inherently unsafe because one drive failing means you lose everything. With RAID 0 you are roughly "n" time more likely to lose everything, where "n" equals the number of drives in your RAID. See https://en.wikipedia.org/wiki/RAID for detailed info.

    For better speed plus security, common arrays are RAID 5 or RAID 1+0. I've built both.

    A relatively common database server approach is RAID 1 for the system drive and RAID 1+0 for the data drives.

    For my photo/video workstation, I use an SSD for the system drive, a two-disk RAID 0 as a caching disk, and a six-disk RAID 5 (with one hot-swap spare) for data. The RAID 0 and RAID 5 arrays use a separate two CPU RAID card driving Seagate enterprise hard drives. (Some Western Digital disks are not suitable for RAID.) And one fixed and one removable hard drive for backup. All in a full size tower case.

    Dan.
     

Share This Page