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.
cociu - hostsolutions.ro - NETSILVANIA | Move your services!
This discussion has been closed.
Comments
He needs disks first
I'm sure many people share the same sentiment (I do), but with all the crypto bullshit happening I doubt he'll be able to get drives to set up new nodes any time soon (at least without getting charged through the nose for it).
And he can't just wipe and reuse the failed disks willy-nilly either (I'm assuming the drives are physically good, just the raid array is fucked) since some customers will probably shit their pants about their data being lost (although it's their fault for choosing HS as their primary storage and/or not having backups of backups of backups).
Conspiracy Theory time:
@cociu is actually using the HDD's himself for Chia farming.
4800 TB for Chia farming equals approximately $122.33k/month.
Probably 100x the revenue he is getting now from hosting...
Too many pages!
One or both of the guys came to your town for food or a girl.
He'd be dumb not to, but it's not like he could plot them all at once, he coulda kicked people off as he needed the space...
Wait, when have I ever defended cociu? Correcting you for being wrong is not defending him. I've been warning people away from cociu since I started posting on LET. I've argued with jsg about him. It was inevitable this was always going to happen. I just thought sooner.
You've gotta stop sniffing glue. Or being so sensitive.
@cociu has been shown to lie multiple times, including in this thread:
So, he has managed to slander himself, and didn't need the help of the community.
(*Unless, maybe, @cociu is using your IPv5++ 64-bit addressing scheme, which isn't tracked on the IRR.)
if only people - incl. you - grasped that.
A propos: You recently said something along the lines of "someone who is ignorant and careless with language should be expected to be ignorant and careless (in other areas like) his operations too".
My first (not communicated) reaction was "nonsense!" but your statement stayed in my head and popped up now and then and made me think about it and meanwhile I think you were/are right with that statement.
Probably a case of non-techie getting something terribly wrong.
I'm amazed how people take that HE listing seriously, as if one could only have (and route) IPs attributed at HE.
I don't know how many customers and nodes @cociu - or for that matter any provider here - has but I know that the IP ranges listed on bgp.he.net is not the upper limit.
That said I don't care anyway about his number of clients or nodes. What we should care about is how well a provider treats his clients and how well he configures and maintains his nodes.
What cociu did (not even providing a FYI) is a cardinal sin. No respectable provider treats his customers like that. And his explanation/excuse for it is ridiculous because a 2 liner here at LET would have made all the difference and cost no more than 1 minute.
This information is available publicly from public WHOIS data from the RIRs, with some help from a BGP table dump from HE's end since they should have some basis for the whois lookups, instead of a brute-force search through the entire IP (or /24) space.
If cociu is operating his own ASN/IP ranges they should indeed be listed there. This information doesn't come from HE's partnerships or some kind of proprietary research method that's also incomplete.
cociu's alter ego is very funny to read, always trying to be right
Exactly.
You can literally download global routing table snapshots and see what ASN is announcing what prefixes at that point in time.
There's an opensource project that I'm pulling a blank on that literally keeps these snapshots public for you to download and parse yourself.
What HE is showing is a snapshot of the global routing table at whatever time they processed it at.
Francisco
https://bgp.potaroo.net/index-cidr.html
http://thyme.apnic.net/
Although afaik HE uses their own BGP table.
That, but there's one that has tarballs of the bgp table dumps timestamped.
Francisco
Isolario shut down.
http://archive.routeviews.org/
?
Guys, keep topic derail within sane limits. Does anyone has more than few servers with kokio?
Lol, around 20 storage servers from me. I don’t think anyone uses their server(s) for production though.
Wow, all servers for iso seeding or anything else? The damage for you is significant.
Correct me if I'm wrong, but he offers Shared hosting. Why would he need 3000 IPs for 3000+ customers?
For example, my VM also affected, and it's not a 'storage VPS' plan. It's an older 7EUR/year, standard, promo VPS. It was originally OVZ, but recreated to KVM (in august, 2020). (The control panel define 'HDD' in the plan name, so it's probably not SSD or other storage solution).
@stevewatson301 @Francisco
Re bgp.he.net: that's all nice and dandy but it's not the complete story. For example I know of multiple providers using multiple entities to hold IP ranges, so while HE's list for HS/NS is probably right @cociu could hold say multiple other ranges listed under a different name. Another point is that HE is about actually routed ranges and it's not uncommon for providers to quietly hold not yet used reserves.
TL;DR: the HE listing is no solid basis to say that some provider only has the listed IP ranges.
TL;DR 2: "cociu only has less than 3000 IPs" was a side story anyway and one that is irrelevant for the topic of this thread.
Re this threads topic I understand those who want to see cociu punished or even going belly up and I'm p_ssed off too and angered by his ignorance and lack of even minimal communication, but I think we also need to see that at least a considerable part of his customers want him to survive and to keep their cheap VPSs.
Bam!
Francisco
OK, so lets get to the meat of it then. Do you truly believe a 'UPS failure' fried 400 drives? You come off pretty smart, there's simply no way you believe it. No one does.
Here's my take:
When you run out of space on a thin provision LVM pool, you can't actually start the pool unless you have at least a tiny bit (even a couple GB) of space spare on the volgroup for the metadata. Without that, it won't start. Infact, for the longest time, the volgroup would destroy itself (eat its own metadata partition) when it ran out of space. If he's on an old enough kernel/lvm version, could've happened.
#2: - He owes his datacenter a lot of money as has been mentioned in the filings and the DC simply cut some of his power to slow down the debt from growing much more.
#3: - A crack pot one, and more so a joke than anything, he sold the drives.
#4: - The outright lie - a UPS fried it.
Lets dive into #4 a bit. A datacenter known as WebNX had a generator (that was indoors) catch fire, set off the sprinklers that were not only above the generators, but also above the servers. The sprinklers soaked the servers full on. Not just a sprinkle. I literally went to the datacenter to help a friend of mine recover his data and the drives all have massive water marks on them.
Even still, they have 1000's, maybe even 10's of thousands, of servers online and they didn't lose 400 drives in their mess. They lost some, but I'd be surprised if it was 100.
I've had nodes go offline because of a UPS on the fritz and it wasn't double cabled. Even still, not a single drive was physically damaged.
If there was such a catastrophic UPS failure in his datacenter we'd have heard news of someone else getting affected by it. There's not a single host that's even heard such a thing as a rumor, never mind actually experiencing it.
I think it's #1 or #2, maybe both.
He didn't have a UPS failure that fried 400 drives.
And remember. Any power strip that'll be allowed to be used in a datacenter has surge protection in them with built in breakers. Those things would've tripped and protected the gear.
Infact, don't APC/Tripp include insurance with every strip for this very event?
Francisco
So, having run out of arguments, are you saying that cociu is using IP ranges not routed on the internet to service their customers?
This jsg guy never missed a single hostsolutions shit show. I knew he would be summoned when I saw this thread days ago.
Chinese call those paid posters hired by the government 50 cents army as rumors are they are paid by 50 cents for every post. I call this jsg guy sisters army, as I speculate he is paid with sisters.
#5 - consumer drives and and most 8TB+ for are SMR. Then he discovered how much fun a raid array rebuild can be...
Very much doubt he used SMR drives. Maybe I'm crazy but i'm pretty sure the drives would've fallen out of any array he put them in (be it hardware or soft/zfs) if they were.
Francisco
This might explain why @jsg is so keen to defend HS. If I had 8 free VPS from a provider, I would damn well back them up in situations like this since I wouldn't have much to lose other than my free service.