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
Reminds me of Hostodo and my 1gb ram VPS that never exceded 32mb. And support that either deleted or closed all tickets when I provided proof of issues.
That's all most people need tbh, just some room to burst. Let's be honest, most people are out there running a Wordpress blog. The upper end of their CPU exists for when their caching plugin purges cache right before Google starts crawling the blog :P
Pretty safe bet for that person that not everyone on the node experiences that same event at once. That's why you can reduce cost by taking on a very reasonable and consistently safe risk.
I've had a huge increase in people trying to signup with me to mine coins and such. The amount of people that have tried to buy 50+ 128MB plans to mine is shocking.
Francisco
Sure, but where do we set the bar? That's all I'm trying to ask here- obviously I'm not really happy with how this has turned out with this provider, and will be unlikely to use them in the future- but how am I in the wrong here by wanting to know exactly what constitutes my fault in this?
Again, I think you're seeing this as an exposed host CPU, rather than a vCPU. As CC knows, you can easily split a single 2.4Ghz E3 thread into 4 vCPUs and oversell far beyond that.
If I wanted to and could afford to pay for a managed, dedicated service, well, I would. I don't need that- It's supposed to be my job to manage my hardware and virtualized setup, too. I'm here for the cheap shit expecting that I can manage the majority of issues myself.
..and where are we now at 100% resource utilization- even virtualized resource utilization? Again, by my metrics only, I had a load of 0.25 in this instance, and it got killed. Maybe we need to stop discussing numbers, because we have a rather wide berth beyond what I use as an example, and yours.
Ugh. Well, we haven't explicitly banned that like we have shoes. Yet. Is it time for yet another "don't do this" that needs to be explicitly stated, when both miners and shoes should fall under "Don't be a dick"?
I was going to wish you welcome (back) to Slackware, but I guess no such luck. :-(
The installer has changed in small ways since 1996 (e.g., ext4 is now the default file system, floppies are no longer supported), but yes, it's basically the same design. This said, it works!
What's wrong with lilo? It works and is much simpler than grub2 (though grub2 is more powerful/flexible). But just to point out that grub2 is included in the A (base) series since 14.1 for those who want/need it. The only catch is that the installer doesn't offer a choice between lilo and grub2, so you have to configure grub2 after the first reboot.
That reminds me of someone who bought 15 Kimsufi dedi during the last flash deal, and people who tried to grab cheap server from Hetzner auction.
Mining has spoilt GPU market, and now even server market? smh
Since this thread is hopelessly derailed into what we define
sexnumbers as, anyhow..I'm a bit surprised to find that UMSDOS isn't still an option for a filesystem.
The way it actually works. The way you have to rewrite it so it knows where to find the kernel every update, lack of filesystem support/etc. I'm not a huge fan of grub, but LILO is pretty funky to rely on in the last decade of EFI and many not-0x80-emulated disks.
Monero is the common one but there seems to be a growing push in the shitcoin market to make CPU based currencies and to make them GPU/ASIC resistant.
Still, If you have 100 (good) cores running full blast, you'd pull $500 -$600/month for a $200/month investment. If it was allowed.
Francisco
If you're not being a dick, experiencing a bug that creates significant unintended strain on the server, or under attack due to a public facing service on your server, the provider should be at fault. I like this better than numbers because it really gets to the core of what the VPS was for and how the risk is managed when done right. It takes the responsibility off of the customer, which is what stated resource guidelines do (pass all responsibility to customer), and forces the host to actual be a host.
You (provider) manage the risk by assuming that legitimate usage, where not being a dick (coin mining on shared resource, eating up 100% of resource until you decide you're done, is being a dick), never equals out to a problem. You assist that by making your hardware, pricing, and resource allocations unattractive to someone who wants to be a dick (might as well just say "mine coins" at this point).
Then there's the node management side of it. This is what I always called "context based administration" at Catalyst. With OpenVZ it is/was always known that the processes show in "top" on the node. If the node is hurting and on top is apache, I would be working to give it what it needs. If the node is hurting and "minerd" is on top, someone is getting canned. Context is so much more important than equally distributed numbers. If, as a provider, you're balancing things in such a way that someone can't use the service for an intended function then you're failing.
The first and last time I tried UMSDOS was in 2002 (if I remember well)! I recall that it was a bit slow as a file system. Plus it had a number of restrictions. Couldn't compete with ext2 at all.
I suspect that there are plans to move to grub, which is why it's in the A series. But for simple, legacy-like setups, lilo works fine. If there's an kernel update in Slackware, the installer offers to reinstall lilo at the end of the update, so you don't have to remember to do so yourself.
I am not against cryptocurrency actually, in fact the idea is great. The only problem is people who are so desperate to get money, which defeat the actual purpose of "mining". Some don't even know how it works, just download some binaries and run it.
Off-topic: I will try BuyVM one day, I had known it since 2011 and remember the good old days where I always refresh doesbuyvmhavestock because the smallest package is always oos (especially the one with Windows is cheap) haha, nice to see it being healthy and respectable until today
Just click this and tell me you can hold out longer than today: https://buyvm.net/kvm-dedicated-server-slices
There needs to be a clause for legal castration for miners on shared resources. Using rusty, dull scissors.
AWS' Burstable Performance (XEN) Instances use the well-known CPU Credit Balance:
>
In my experience, AWS and some XEN VPS providers sell very low baseline performance VMs but supersize the initial CPU Credit, giving users a "feel good" first impression. For users running just a couple of tests and benchmarks after the deployment the VPS looks fine but the actual performance may be shocking for extended usage.
*Ahem*
FUCK BEZOS!
KVM Babay!
@WSS
Most shitty OVZ providers that limit CPU define it as 20% of a thread sustained. Which, IMO, is worthy of:
#DICKS
You can, and honestly, in the real world, no one will care about 18% but if you are using a sustained amount that is not what you are paying for (probably)
That's when it's bloody running tho..
@AnthonySmith
well if that's only an hour or so then I would call you an ideal customer.
I told you to wait for the Slabs to be available before you starting your shilling run.
JEEZ.
Francisco
Basically most (if not all) threads end up being a marketing thread for @francisco
This thread is confusing... Is the supposed 'limit':
0.25 loadavg (sustained).
OR
0.25 CPU usage (sustained).
There is a big difference. Figure it out.
And that's all I have to say about that.
He deserves all the shilling we do for him.
I mean, if you want to be on the payroll just let me know.
Francisco
Thanks for dropping by.
I'd be pissed too.