[nik736]

It's an interesting article, thanks for that.

What people forget about the OVH or Hetzner comparison is that for those entry servers they are known for, think the Advance line with OVH or AX with Hetzner. Those boxes come with some drawbacks.

The OVH Advance line for example comes without ECC memory, in a server, that might host databases. It's a disaster waiting to happen. There is no option to add ECC memory with the Advance line, so you have to use Scale or High Grade servers, which are far from "affordable".

Hetzner per default comes with a single PSU, a single uplink. Yes, if nothing happens this is probably fine, but if you need a reliable private network or 10G this will cost extra.

[lossolo]

These concerns are exaggerated. I've been running on Hetzner, OVH and friends for 20 years. During that time I've had only two issues, one about 15 years ago when a PSU failed on one of the servers, and another a few years ago when an OVH data center caught fire and one of the servers went down. There have been no other hardware issues. YMMV.

[hedora]

They matter at scale, where 1% issues end up happening on a daily or weekly basis.

For a startup with one rack in each of two data centers, it’s probably fine. You’ll end up testing failover a bit more, but you’ll need that if you scale anyway.

If it’s for some back office thing that will never have any load, and must not permanently fail (eg payroll), maybe just slap it on an EC2 VM and enable off-site backup / ransomware protection.

[ghaff]

Wasn't my product as a product manager but my long-ago company came out with an under the desk minicomputer product for distributed sites. And they didn't use ECC memory in the design. The servers didn't fail very often but multiply that fairly low error rate by a large number of servers and a system failure was happening every few days or so. The customer wasn't happy.

[torginus]

I never understood the draw of 'server-grade hardware'. Consumer hardware fails rarely enough that you could 2x your infra and still be paying less.

[jammo]

Yes, but there are options for dedicated server providers who offer dual PSU and ECC ram etc. It's more expensive though for e.g a 24 Core Epyc with 384GB RAM dual 10G netowork is like $500/month (though there's smaller servers on serversearcher.com for other examples)

[montecarl]

I can't believe how affordable Hetzner is. I just rented a bare metal 48 core AMD EPYC 9454P with 256 GB of ram and two 2 TB NVME ssds for $200/month (or $0.37 per hour). Its hard to directly compare with AWS, but I think its about 10x cheaper.

[titanomachy]

Wow. Probably performs better too, with a recent CPU and non-"elastic" disk. What about ingress/egress?

[wombarly]

Hetzner dedicated servers by default have 1 GBit and free egress. You can opt for 10 GBit, which has 20TB and then €1 / TB overage.

[hedora]

Their current advance offerings use AMD EPYC 4004 with on-die ECC. I can’t figure out if it’s “real” single correction double detection, or if the data lines between the processor and dimms are protected or not though.

[nik736]

It's only on-die ECC not real ECC

[vjerancrnjak]

Is there software that works without ECC RAM ? I think most popular databases just assume memory never corrupts .

[torginus]

I'm pretty sure they keep internal internal checksums at various points to make sure the data on disk is intact - so does the filesystem, I think they can catch when memory corruption occurs, and can roll back to a consistent state (you still get some data loss).

But imo, systems like these (like the ones handling bank transaction), should have a degree of resiliency to this kind of failure, as any hw or sw problem can cause something similar.