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
We did it boys. two hunnit pages of crying about a $5 server.
Oh, Octave sended a DM to me lmao about my KS-A order
6 inch or 12 inch baguette?
ahem, ahem that was my idea! I Originally ordered ks-a through a selenium bot
I have this server for transfer (CA account)
KS-LE-1 GRA (upgraded ram and new disks from 2x480 GB to 2x960) - Renewal: $11.05/month
Im sure you're getting it as well, i guess im not gonna get reached.
I don’t see why they don't make them money. Basically, they stuff "thin-client" level hardware with a bit more memory in racks. Something like GIGABYTE GA-H110M-A, Dell XPS IPSKL-VM or some BMC crap.
Retail price for systems like usually cost like ~500$ per unit retail. Most likely they got it with 20-30% discount. They rented them out for 129$ back in '17. My bet is that they got all their CAPEXes back in the first 18 months. So for the rest of their lifetime they cost them power, connectivity and floor space == peanuts.
So yeah, I'm absolutely sure they are still making them monies
Let's see tomorrow what will happen
Bad PR is when your burning datacenter is uploading client's data to the literal cloud. That's like really bad-bad PR. But hey, do you remember that? Do you like reaaaaaaally care? Nope. We all here still sniping outdated hardware for cheap.
IMO, this one is quite good PR. 200 pages and counting worth of good PR
Even ignoring HW, no way €5 pays for electricity, network, cooling, the physical space that it takes up AND support.
It’s for marketing.
It’s clear since other WORSE servers in the Kimsufi line cost more, as well as by its unique naming scheme.
kimsufi servers are less reliable than other Rise, Sys plan. it is always recommended to have a offshore backup of the data. this apply to all Kimsufi range.
Even a brand spanking new server could die
Well, that applies to any host THB. Overall KS servers give you a lot of bang for your buck.
Yes, you are right. this means many new users can buy it and experiment with their projects. kimsufi range are a great options to start with. and as always, do backup. every single time.
Did someone say spanking
@ehab !!!
Especially disks are in near death experience. Got a KS-2, 4x2TB disks, had to change 2 out of 4. One of them did not sync at normal SATA speed, and the other one gave constant checksum errors on zfs raid scrub, even though SMART was ok on both of them.
i assume sys and rise plans gets into kimsufi range after few years of uses and they replace higher plan with newer hardware. that could be the main reason kimsufi faces issue with hardwares.
Yes, that's basically their business plan. Truth be told though, they have always replaced failing hardware quite quickly in my experience.
They made Powerpoint worthy diagrams of this: https://eco.ovhcloud.com/en/about/
Even the title of the chapter is Powerpoint material: Competitive pricing for a responsible provider
yes they are quick to replace failed hardwares.
that's the transparent, easy to understand page for kimsufi servers.
Welcome to magic lang of economy of scales. Math is weird here. This folks have a bloody factory to build custom hardware for this single purpose: make monies on scale. So let's look at money:
FY24:
Revenue: €993m, w/gross margin 38% (cloud is money printing machine)
COGS + OPEX = €336m this is how much it cost to run whole empire day-to-day.
Sadly, they don't publish total area or capacity of all 43 DC's.
But they say BHS+GRA+RBX=896k servers. So gross opex even for just 3 of 43 gives us 375€/yr or €31/mo per server.
Marketing? Sure. Have they lost even a single penny on that? Hiiiiighly unlikely.
Considering that the storage devices are "old", what do you think is the "safer", 3 x SSD or 3 x HDD (in RAID5) ?
Nothing is safe.
If you care about your data you will not YOLO and simply have enough backups/ DR plans available.
But yeah, I would go with RAID0, No backups and live one the bleeding edge
Has anyone ever recovered from a disk failure in a RAID5? By recovered I mean lived long enough to see the end of a rebuild after changing one disk?
And boy they do. And what’s weird, folks I’m consulting had pretty shitty experience when they got one of few “high grade” servers DOA and waited for a week or so for replacement, being ghosted by support in meantime. Not a kind of treatment you expect when you spend half a grand per month per server.
What I care the most is not having to use my backup and having to reinstall everything
This is why I was wondering if SSD or HDD made a difference regarding the risk of failure.
Yup. Raid0 is the way to go.