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.
Comments
Bad news everyone, it has been delayed.
Monday, 22 March is it for now.
https://www.ovh.ie/news/press/cpl1786.fire-our-strasbourg-site
It is unlikely that OVH has a control over how fast they can turn stuff on.
Inspections need to be made by the fire department and their insurance company to deal with, not to mention their power company calling the shots.
OVH is the epic winner of this fire. Oles is such a chad. I've been so impressed with their handling of this situation that I bought 5 more servers with them. They'll make more money than ever. Real suckers are people screaming on Twitter that had no backups of their CrItIcAl dAtA.
As for the reason... data centers don't usually catch fire, right? Right?? I think they had data that somebody really wanted gone, so the entire DC had to be taken down. Sabotage is the only plausible explanation at this point. I think further down the line it'll become very clear this was what happened here.
Rigging up the site for a massive UPS/battery bank/generator explosion would've been far more believable than a fire, but would've left no survivors. The staff should be really grateful.
And remember. Today it's OVH, tomorrow it might be your data center. Once somebody is determined enough, they're able to accomplish anything.
Ye yee... these bitches deserved a good burnin'
I do recall hearing about DC fires occasionally but, as far as I know, none of incidents I've heard was able to spread, probably due to concrete floors.
This hashtag is a goldmine
Sadly, some of these people are clearly trolls, just joined this month.
Most people are demanding access to their burnt data.
Meanwhile, some are asking for the data to be re-deleted:
OVH has put a status page where you can check if your service is recoverable:
https://www.ovhcloud.com/en/lp/status-services-backup-strasbourg/
Most interesting thing from my perspective is is that it seems that lost almost all data from "cloud archive" service, aka long term storage.
Ooh... Ooh... double burn
Hello,
looking for someone affected by the accident, who had backups anywhere, could start a new server in another DC and used their FAILOVER IP to successfully move it to this new server.
So I know if this service (failover IP) is reliable for situations like this.
Thanks.
According to Twitter, users were receiving error messages trying to move a failover IP out of SBX. Some users tried for many hours and finally got through.
It would be faster to switch the DNS records instead.
We have it better (as we do storage replication on ~5 minutes between different OVH datacenters) + cold reserve (storage replicated on one hour to another provider) + backups BUT
Failover IP movement function get clogged in that case and trough the day they have service degradation (of the movement) in entire EU
We started IP movement from SBG to GRA in 4:40 AM CET and it took more than 6 maybe 7 hours - I was ready to give up and start the VMs with new IPs
then trough the day I started to move another IP i do not use between GRA and RBX (timing dropped to 2-3 hours) but was slow... In the evening it was back to normal to 3-5 min
They claimed that they added capacity to that service so that will not happened again
You wouldn't have guessed what message I got popped-up when entering OVH website..
The protection of your data is our priority.
Well, to be fair, they did send customers' data to clouds, actual clouds.
Now, that is pure dedication on their part. It must have cost them a lot, too. Show me one provider who does that for ya, none.
OVH forever.
First real cloud provider. No one promised that data will be retrievable.
Those data in the clouds won't ever be breached. That's the best security.
Its data in the smoke not in the clouds.
Meh, that's a bureaucratic choice of words.
Smokes, clouds, they are all tiny particles of something in the air. To OVH, they are the same.
i thought some colors to
https://www.ovhcloud.com/en/lp/status-services-backup-strasbourg/
would be beneficial so i give you this little chromeconsole snippet
$('td').each((k,v) => { v = $(v); switch (v.text()) { case "Recoverable": v.css("background-color", "#00ee0099"); break; case "Non-recoverable": v.css("background-color", "#ee000099"); break; case "Under investigation": v.css("background-color", "#eeee0099"); break; }});
note the many red with openstack zone 1,2,4...
Thanks for sharing.. so it seems one should go directly with IP change with such an event, it can be predicted that the service will degrade.
By the way how do you replicate the storage every 5 minutes between DCs?
I hope they will fix it cause there is no point of having failover IPs when you cannot failover in time and they adverse them for that at least in the past ...
For one domain dns failover is ok but for hundreds when half of them do not use your dns becuse in forum told them no to use providers dns it become complicated
Proxmox ZFS storage replication https://pve.proxmox.com/wiki/Storage_Replication
after first sync it is super efficient and took seconds to sync changes
Not sure if I understand the correctly, but what are "Internal Backup (DC/Service)", e.g. "Public Cloud Archive (os-pca-sbg)" that are in the same rows as VPS zones? Does it mean they replicated the data from the VPSes if you didn't pay for it or selected any backup options? Or are those for paid Public Cloud Archive option. Mostly concerned about what "Internal" means in this case.
I was never particularly worried about backups, but always concerned if data actually disappears from existence the moment I delete it from a VPS. This situation of non-recovery seems to confirm it luckily, in case of OVH at least.
Thanks! While sync is fast, is there any reliable way to check that the copy is correct and consistent?
Some years ago I've found a bug in Bacula that was causing some percent of the files to be actually missing in the backups; since that I'm really paranoid about this. I was curious how many people at that time were not aware that they are doing backups that are useless
Bacula & ZFS is not comparable as ZFS has way more users starting from Data Hoarders with Free Nas, Rsync.net, Hetzner storage boxes and Other big biz... (Open ZFS I mean, non open one is hoarded by Oracle)
Can bug happens? Yes it can .... As it happens as happens 6 years ago... https://github.com/openzfs/zfs/issues/4050
Can a bug happens before the community finds it and simultaneously in same moment the main server burns , the hot spare got broken file (not entire one file system but a file) by a zfs bug the cold storage (could be affected with the exact the exact same file) and on top of that regular rsync jetbackup can be failed in exact same moment
Yes It can...
But then it will be act of God and will be very hard to sue them and prove negligence from your 2.95/mo shared hosting provider if they take all that steps to protect your data
Jesus Christ, you went 0/2. They specifically DON'T want ANYONE to access it and they want it DESTROYED, not re-deleted. The tweets were pretty clear.
It's perfectly valid to prevent damaged hard drives from being sent to the recyclers and then had their data read (and information misused). The fact that you're Chinese and actively dismissing this valid scenario makes you super suspicious.
SMH
"re-delete" and "destroy" mean the same:
I run this when I need to return a disk to the supplier.
They don't, don't be obtuse. There will be zero documents on hard drive shredding from any company that includes that dd command. It's also ridiculous to ask OVH to run dd on the pile. Right, because nobody fucking did.
Re-delete and destroy are not interchangeable in any use case where someone really does mean "re-delete", such as editing a document.
@yoursunny was simply pointing out an interesting contrast. No need to make the miscommunication a huge deal.
encrochat was hosted at ovh ^^