Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


Dr. Server RAID Failure - All Data Lost
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Dr. Server RAID Failure - All Data Lost

Anyone else get this?

`Dear customers,
We have suffered 2 drive failure on the hypervisor hosting your VPS. This has resulted in the RAID array completely gone, thus destroying all data on it. We do extremely apologize for the inconvenience caused, but there is nothing we can really do at this stage. We can offer a full refund of your last monthly payment(pro-rated for yearly payments) or a clean VPS with 1 month compensation. How to proceed?

Please open a ticket at your respective billing system: https://portal.drserver.net or https://vikinglayer.com/clients to inform us of your decision.

Once again, we do extremely apologize for the inconvenience. Such events do happen. They are rare, but sadly they exist and we have been hit by one.

Thank you for your understanding,
drServer.net/VikingLayer.com`

«134

Comments

  • Ouch!

  • shit happen.

  • I wonder what the array size was, what size and brand of drives, whether a recovery was in progress when the 2nd drive failed, etc.

    Thanked by 1Janevski
  • @budi1413 said:
    shit happen.

    Yea, just sucks as I was doing a slow migration of my sites over to there and now I gotta start over.

  • @willie said:
    I wonder what the array size was, what size and brand of drives, whether a recovery was in progress when the 2nd drive failed, etc.

    Not sure, I did open a ticket about random IO speeds, one hour it'll crawl at 10 MB/s in another hour it'll be over 150 MB/s

  • LeeLee Veteran

    I minor inconvenience, everyone had their own backups, right?

  • deankdeank Member, Troll

    Back up is for pussies. Real men don't keep back up. Real men live dangerously.

  • winnervpswinnervps Member, Host Rep
    edited August 2018

    @deank said:
    Back up is for pussies. Real men don't keep back up. Real men live dangerously.

    The end is nigh, right?

  • winnervpswinnervps Member, Host Rep

    @Nomad said:
    End is nigh for pussies. Real men don't end.

    Real mean eat pussies

  • @winnervps said:

    @Nomad said:
    End is nigh for pussies. Real men don't end.

    Real mean eat pussies

    real man eat ass

  • Are they really using BoxBilling as client area?

  • deankdeank Member, Troll

    What I've learned from this thread.

    Real man eat something.

  • deank said: What I've learned from this thread. Real man eat something.

    And won't keep backup, because they should live dangerously.

    Wow, so much learning! This is not LET, it is Wikipedia!

    On a serious note now: Woops! Shit happens, but it is always frustrating for the clients. Backups or not, it is downtime + effort to restore... Even to really good providers like Andrej @DrServer and @Radi ...

  • jsgjsg Member, Resident Benchmarker

    @Lee said:
    I minor inconvenience, everyone had their own backups, right?

    I guess not. Sounds like a case of "backup not needed. We have Raid drives".

  • jsg said: I guess not. Sounds like a case of "backup not needed. We have Raid drives".

    Raid is not backup. Anyone who thinks that raid is a replacement for own backups, really likes vivere pericolosamente...

    Thanked by 1WebProject
  • deankdeank Member, Troll

    Anyone who knows about computers a little knows raid is never a backup solution.

    But then even Linus thought raid was a backup solution before his raid got corrupted.

  • EHRAEHRA Member

    Such a situation is terrible, I hope I never have to go through it. I hope all those affected have up-to-date backups ...

    image

  • 2-drive failures aren't that common but they happen. I wonder what the raid level was of this system (forgot to ask earlier) and what kinds of services it ran.

    Absolutism about backups or anything else isn't helpful. If someone is injured in a car accident we don't say they should have chosen to drive a tank instead of a car. Rather, risk assessment is a thing and the best we can do is make informed decisions when possible.

    Thanked by 2angstrom zed
  • WebProjectWebProject Host Rep, Veteran

    @Lee said:
    I minor inconvenience, everyone had their own backups, right?

    haha, I bet you that not everyone as some people relay on RAID

  • @deank said:
    Back up is for pussies. Real men don't keep back up. Real men live dangerously.

    Do you always use pull out method?

    Thanked by 2agonyzt PrestigeWS
  • deankdeank Member, Troll

    If must use raid, real man use Raid-0. Why zero? Cuz zero sounds cool AF.

    Thanked by 1WebProject
  • edited August 2018

    Hate to be a monday morning quarterback but I have seen this happen before. RAID is useless if you don't set up automatic notifications of a drive failure. Usually, the drives have similar hours on them so fail around the same time. If one drive fails the other has to work harder and then it fails not too long after.

  • AnthonySmithAnthonySmith Member, Patron Provider

    Sometimes its not so much a physical drive fail, I have had 2 x hp 410i's throw out 3 drives from a raid 10, raid is good for redundancy but with HW Raid you still have that single point of potential failure.

    1st time I put it down to a freak accident, 2nd time I was gob smacked, managed to recover the data on both occasions although it took so long I wished I had just declared a data loss and insisted on restore from backups.

    These days with most things moving to SSD dealing with large volumes is much less painful.

  • raindog308raindog308 Administrator, Veteran

    deank said: f must use raid, real man use Raid-0.

    YOLORAID!

    lemon said: real man eat ass

    jvnadr said: On a serious note now: Woops! Shit happens,

  • williewillie Member
    edited August 2018

    LosPollosHermanos said: If one drive fails the other has to work harder and then it fails not too long after.

    It also might be preferable to take the server offline during the rebuild, rather than trying to have it serve user traffic at the same time as rebuilding. That speeds up the rebuild a lot and keeps the i/o sequential, decreasing risk of 2nd failure. Finally if the drives were near EOL when the first one failed, perhaps both should be replaced. I don't know that 2-drive HDD RAID-1 is still a thing for normal VPS host nodes any more though.

  • jsgjsg Member, Resident Benchmarker

    @LosPollosHermanos said:
    If one drive fails the other has to work harder and then it fails not too long after.

    ?? please elaborate.

  • deankdeank Member, Troll

    Well, in a proper raid 5 or 10 or something of that sort, a single HDD failure doesn't bring down the entire raid and operation can go as normal.

    It also means other drives must work harder due to absence of the dead drive. The extra sustained pressure sometimes bring down another HDD that was dying already.

    Raid drives tend to come from the same batch, so when something fails, the others tend to fail at a similar moment.

  • jsg said: ?? please elaborate.

    In a 2 drive RAID system the read load is split across both drives, so if one drive fails the other has to serve all the reads => higher load. Very high load also results from a rebuild in progress, but that's a separate situation from a failure not being noticed soon enough.

  • jsgjsg Member, Resident Benchmarker

    @willie said:

    jsg said: ?? please elaborate.

    In a 2 drive RAID system the read load is split across both drives, so if one drive fails the other has to serve all the reads => higher load. Very high load also results from a rebuild in progress, but that's a separate situation from a failure not being noticed soon enough.

    That's only the case when striping and even then the details depend on the implementation.

    When mirroring the remaining drive does NOT have to work harder.

Sign In or Register to comment.