Howdy, Stranger!

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


BuyVM: All data lost on nodes 27, 41, 52, 58 - Page 3
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.

BuyVM: All data lost on nodes 27, 41, 52, 58

1356

Comments

  • yomeroyomero Member
    edited May 2012

    @Francisco said: You were on node27 i think.

    Indeed

    Sent to "Backups/Restores"

    Probably the wrong department lol :$

    Ty Fran

  • manmamanma Member

    I made a ticket early on in the day, and I've since replied with a few directory names, but nothing's been done as of yet. Is node 41 still being FSCK'd, or was I supposed to open a new ticket or something?

  • RophRoph Member

    Opened ticket too, everything recoverable would be nice from my 128M but all that's really important are my ZNC configs. And manma, just give them some time :)

  • manmamanma Member

    @Roph said: Opened ticket too, everything recoverable would be nice from my 128M but all that's really important are my ZNC configs. And manma, just give them some time :)

    No worries. I can as long as I need to. I just don't want to get accidentally passed over because I didn't ask the right way or at the right time.

  • FranciscoFrancisco Top Host, Host Rep, Veteran

    41 is online as far as i know.

    Anthony is chewing through tickets as he can and I'm helping him with any VM's he needs help with.

    It's quite possible he's sleeping right now and will be back in a bit :)

    Francisco

  • tuxtux Member

    @Francisco said: It's quite possible he's sleeping right now and will be back in a bit :)

    He sleep at work?

  • FTNEthanFTNEthan Member
    edited May 2012

    @tux -

    He needs energy in order to work efficiently, I imagine it must take a lot of hard work to deal with 4 different nodes with lost/corrupt data.

    and its midnight in the US.

  • klikliklikli Member

    @Francisco said: We've found ways to start identifying CT's but please, if you're able to easily just rebuild let us know as that'd help a lot.

    Seriously, what should I do next? Should I wait or do anything else?
    Concerning the last sentence, yes, you could probably just rebuild mine as all it got is a blank installation of NGINX. My VPS ID with Stallion is #2862.

  • FranciscoFrancisco Top Host, Host Rep, Veteran

    @klikli said: Seriously, what should I do next? Should I wait or do anything else?

    If you are able to rebuild easily (read: you can remake conf/files/etc very easily) then it'd be a huge help to us if you're able to just do that. Log a ticket and let Anthony know so he can cross you off the list.

    If there's files that you could really use back we can see if we can dig them out of our cold, dead, nodes.

    Francisco

  • KuJoeKuJoe Member, Host Rep

    This thread brings back memories...

  • CoreyCorey Member

    Whoever said 'this is reason to use software raid' - is an idiot..... This catastrophe would have been just as bad on software raid because a hard power down will always result in data loss on software raid.

  • pcanpcan Member

    I had a VPS on node 27. The fastest route to solve the issue is to use the BuyVm ticketing system, as Francisco suggested multiple times already. The BuyVM service has been well above my expectations. I timely received the e-mail (listed at the beginning of this tread) that fully and honestly disclosed the extent of the issue, so I immediately switched live service to a spare machine. Then I opened a ticket to ask about recovery options; Anthony almost immediately answered that data could have been corrupted, so I decided to restore from my backup. A new blank VPS was available in a very short time (I was not expecting such a fast response), and I restored the backup in few minutes.
    I see that Francisco is now trying to recover data to people without backup. This is truly beyond my expected service level for a 15 USD/year unmanaged VPS, and I really don't understand some angry comments here.

    Thanked by 1Francisco
  • CoreyCorey Member

    I agree with you pcan.

    1) Keep backups
    2) Don't expect your unmanaged provider to have backups for you or restore your data if something happens to the node out of their control.

  • FranciscoFrancisco Top Host, Host Rep, Veteran

    @Corey said: hard power down will always result in data loss on software raid.

    As well as a full data resync :/

    I don't think any of the angry people are customers, just people trying to troll @Aldryic. With stallion I quickly wrote a mass reprovision script and Anthony had full nodes rebuilt in a matter of minutes. A few CT's didn't get made but that's because there was some outdated template data for them.

    Francisco

  • FranciscoFrancisco Top Host, Host Rep, Veteran
    edited May 2012

    @Corey said: 1) Keep backups

    2) Don't expect your unmanaged provider to have backups for you or restore your data if something happens to the node out of their control.

    We do include 5GB of space and are ramping it up to 15GB for everyone in the coming week or two as I said. While 5GB might seem a little lacking, for $15/y VM's this is likely enough for most people to store everything on their VM's.

    With the ramp up to 15GB and the option for people to justify more if they need it, there's going to be little excuse for people to not use the service outside of them liking to make things complicated for themselves.

    Francisco

  • CoreyCorey Member

    @Francisco - we don't usually think of our power strips being a POF since they are usually pretty simple in design - not much you can do to make this redundant except have dual power supplies on two different power strips ran from an A and B source.... but running A+B power on budget boxes is really expensive.

  • CoreyCorey Member

    We do include 5GB of space and are ramping it up to 15GB for everyone in the coming week or two as I said. While 5GB might seem a little lacking, for $15/y VM's this is likely enough for most people to store everything on their VM's.

    With the ramp up to 15GB and the option for people to justify more if they need it, there's going to be little excuse for people to not use the service outside of them liking to make things complicated for themselves.
    Francisco

    @Francisco - I know, while in theory this should keep you from loosing some customers.... you're always going to have those few that don't understand and think you are a terrible provider. I hate the way that works.

  • FranciscoFrancisco Top Host, Host Rep, Veteran

    @Corey said: @Francisco - I know, while in theory this should keep you from loosing some customers.... you're always going to have those few that don't understand and think you are a terrible provider. I hate the way that works.

    We've had no cancellations on this that i'm aware of. A+B isn't possible in our racks as coresite won't do 80A to a rack and us only have 20A of sellable servers in a single rack is a real waste of cash.

    If some are upset then all I can do is apologize and work to provide additional services like the improved storage. FSCK's happen to all hosts at some point, I just hope it's minor stuff and it addresses itself quickly w/o any moved files.

    Francisco

  • CoreyCorey Member

    @Francisco yea - it would more than double the cost for having the clients. You did a great job with the recovery.

  • pcanpcan Member

    @Francisco: dual power supply with dual power cords are still a useful redundancy. Price may be lower now. If I correctly understand, on the latest Prometeus offer, a server with dual power supplies is listed as host. But on some services, I still choose lower price over more redundancy (the real redundancy is a second independent VPS).

  • OliverOliver Member, Host Rep

    @Corey said: Whoever said 'this is reason to use software raid' - is an idiot..... This catastrophe would have been just as bad on software raid because a hard power down will always result in data loss on software raid.

    Really? I've had many hard power downs with mdadm RAID1 and RAID5 sets and it has certainly not always resulted in data loss.

  • CoreyCorey Member

    @Oliver you aren't running 100's of customers constantly hitting the disks either I assume.

  • CoreyCorey Member

    @pcan - and if you weren't running A+B power how would you switch the power over to the other 'redundant' strip and power supply without it being interrupted?

  • klikliklikli Member

    @Francisco said: Log a ticket

    Done. But it does appears to be "online" at Stallion - does that mean a VPS has already been rebuilt for me actually?

  • CoreyCorey Member

    @klikli yes Francisco said he created a mass rebuild script - so they have already created you a fresh new vps.

  • OliverOliver Member, Host Rep

    @Corey said: @Oliver you aren't running 100's of customers constantly hitting the disks either I assume.

    I wasn't referring to any system in particular. My point is that it won't "always" result in data loss which is what you stated. :-)

  • CoreyCorey Member

    @Oliver well I meant in the lowendbox provider context - sorry :)

  • prometeusprometeus Member, Host Rep

    @Francisco said: We do include 5GB of space and are ramping it up to 15GB for everyone in the coming week or two as I said. While 5GB might seem a little lacking, for $15/y VM's this is likely enough for most people to store everything on their VM's.

    People start thinking about backups only after the second (or third) data loss. We provide free ftp storage on request with our business plans, NO ONE asked for the access to his ftp area out of 50-60 accounts...

  • pcanpcan Member

    @Corey: Dual power supply helped me many times. As example: I experienced power supply unit failures (the surviving power supply avoided downtime) and also accidental power cord unplug (while moving the server on rails).

  • prometeusprometeus Member, Host Rep

    @pcan said: on the latest Prometeus offer, a server with dual power supplies is listed as host.

    We have 16A+16A or 32A+32A to our racks powered independently, so almost all our servers are dual powered, but as I understand buyvm has a single line of power...

Sign In or Register to comment.