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.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
What if your VPS provider HDD gets corrupted?
I have been using different services on LET.
Recently, I faced an issue with one of the providers.
Their storage server HDD got corrupted, and as a result, some users lost their data.
I'm not here to blame anyone.
I'm just curious about the industry standard in situations like this.
What generally happens when a storage server's HDD fails? Should the provider be expected to have restore points/backups, or is data recovery typically the customer's responsibility?
Comments
You get a blank VPS and an SLA credit for lost days.
You can then activate your disaster recovery process.
Isn't a storage server supposed to be more reliable? People usually use them to store backups and other important data.
In a case like this, shouldn't there be some kind of redundancy, restore point, or disaster recovery process on the provider's side?
What a hilarious timing, I did discover that this has happened to one of my servers today. Well, there were more issues than just that.
I like coincidences.
I just restored from a backup, well, actually I'm still restoring but most of it is done
In my case, no clear communication from the provider; I am just keep on waiting; either build a new server or they fix it.
Storage costs money. Some providers offer automated backups, others offer snapshots for you to take manually, others offer block storage as a different product. No matter how you look at it, storage costs money, therefore services like these are either included and you pay extra, or these are features which you order separately. If you don't opt in for any of these because you probably don't need them, that's great, because it means you make backups yourself and don't wish to pay extra for something which you don't need at this provider (whoever this is).
Hi,
you are supposed to get what is written in your contract.
Does your contract include backups of your data taken by the provider? Then the provider must have them.
Does your contract not include it? Then you must have them.
The best choice is to make sure that YOU always have them - no matter if your provider is responsible for it. If the data is important for you, you should have multiple copies and not rely on one ( the provider ).
Imagine the provider deadpool, datacenter catches fire regards to ovh or the dog bite the power cord of everything. Even you might have a contractual right that the provider has to have backups, what do you do if he does not? Your data are gone. No money can bring them back.
So if your data are important, you hast have a backup / restore strategy that you control.
If not, then not... maybe the loss of all those cat pictures is nothing that much bad... ;-)
Every server should ideally be reliable. The term storage doesn't make a huge difference in that regard. It mostly means that there's (usually a lot of) storage. Now if every bit gets mirrored 10 times or if it's just yolo raid... who knows? That's a question of quality, which again is likely a question of price and even the highest possible quality can still fail. In the end storing something in just a single location is always a gamble that might result in loss.
all i got was sorry your data is gone, we can either move you to LocationA or LocationB.
Which makes me wonder if I’m receiving inadequate support because I bought the server at a discounted rate
The truth is data backup is 100% the customer's responsibility. Budget hosts use RAID for uptime, but if the whole filesystem gets corrupted, RAID can't save it. If you check the ToS, providers never guarantee data safety. They’ll just give you a fresh blank server and move on. Always keep an offsite backup. Never trust a host with your only copy
Did you just imply that a respectful LET member can not go naked and enjoy the YOLO motto?
I know what you refer to, what they 'owe' you is just putting back service to working condition again.
Your lost or corrupted data is not their responsibility, it's your call to get daily backups
This damages the trust as in 'if this will happen again or not' with that provider.
But overall, if they just put service to working condition again their responsibilities would be done. Offering SLA or alternative location is just good will gesture.
Depends what they offer for it. I've no clue what the Hostsystem I'm on actually runs, I would suspect zfs but no clue. Certainly was a surprise and that's already a big crux. If they do not advertise like redundant replication, you're looking at keeping your backups yourself 100% of the time, cause hardware faults will just screw you. With something like a ceph cluster, I would start to dare storing without keeping backups.
Now, if you can just rely on s3, you can get somewhat of a piece of mind if you use garage for example for some geo distributed redundant setup, I would dare to trust that. I did try to break it and so far it handled it well.
Hi,
no of course not sir! I would never dare to give any disturbing advices to free people!
Everyone shall enjoy his freedom in choosing his favorite path down to hell
There is a german phrase: "Nur die Harten kommen in den Garten" it means something similar like "Only the toughest shall enter heaven"
Yeah, in those cases (esp on LET, but in general), it depends on how quickly you need it back up and running. Always have a contingency plan.
Also: some providers offer storage server that isn't even RAID-redundant. You need to read the fine print/offer sheet. Those that do, are pretty diligent in communication as they didn't expect a failure. Still, 1-2-3 backup plans are a must if it's important to you.
Easy, you just need another storage server to backup your storage server!
Ask Australia about Google on this one..
Hard drives cost like $30+ per TB these days. If I'm paying between $12 and $24 per TB per year for my storage VPS which is Raid 1 or Raid 10, I know that just the hard drive space is going to cost the provider at minimum one year worth of fees. When you factor in electricity, bandwidth, and the hardware that the drives are connected to, they probably aren't in the black until, what, year 4? If I want a backup on top of that, we're pushing their payback to year 5 or 6, so no, I don't expect the provider to have backups for me unless they advertise it.
Given those numbers, I'm expecting them to use bottom of the barrel used hardware and something is going to fail eventually. I'm not storing anything that can't be recovered from somewhere else on an LET storage VPS. That said, I have had good luck so far with LET storage, but I'm not putting anything valuable on just one.
Grief, accept, move on!
i'm paying exactly $12/TB/yr with @layer7 , what should I do? /s
Uh, cancel it?
Treasure it forever?
I fall to the floor and cry in fetal position.
I've also spread my data over 5 different providers in 4 countries so I won't lose everything as well.
You and me both and I'm keeping that shit until the end of time. There's a reason I shill them.
EDIT: Also, credits to @idroid007 for making a constructive post - not a childish callout post
Backup your servers using 3-2-1. I have 3 different physical locations for backups of my proxmox backups.
I wonder when DeluxHost will realize that their storage is returning I/O errors for quite a few sectors.
I'd open a ticket for them but the VPS is offline so I can't read the badblocks.log file.
Standard practice is best-effort. The provider will try their best to help you, yet they have no obligation beyond that best-effort. The obligation of backups always falls on yourself for unmanaged services.
I've lost servers with Linode, Vultr, DO, AWS, Azure, doesn't really matter here.. data loss is pretty universal, it'll happen anywhere eventually. Always make backups, shit happens and you need to be prepared.
Just consider it a learning experience that RAID is not a backup, and a SAN with redundancy is not a backup, ZFS is not a backup, etc. Tons of causes of data corruption.
EDIT: To be clear though, "IF" it's a managed backup service then the provider is expected to employ good practices on your behalf since that's what you're paying a premium for.
A Storage VM, it's on you.
DediRock lost all my storage twice on a 2 TB storage server on the east coast, and once on a small general-purpose VPS in LA, all within a couple months. They refunded my money for both servers, but it cost me a huge amount of time to deal with these issues. Never going back.
Short answer: It might happen some day.
Long answer: It might happen some day but then the real question if they will do anything about it.
Only thing i trust for backups is my own machines at home.
Certainly not some fucking VPS lmao, anything important should be autobacked up to your local machines
Probably not.
Thank god modern filesystems let you add blocks to a bad blocks blacklist so I won't have to worry about random, silent corruption (at least until the disk failure gets worse, that is).