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
I explained this earlier but I will do it again.
When you buy a more expensive processor, you need to (over)sell more VMs on that CPU, to make the same amount of money.
We will likely update this page sometime soon - https://hosthatch.com/benchmarks since it's from 2017 and no longer as valid as it was then.....but you get the idea.
I can only say this: if you care about seeing that latest EPYC or Gold in your /proc/cpuinfo, then sure, we lose. But if you care about actual real, consistent CPU performance (that you can measure by GB4/5 for example), then we will likely win when you run the same benchmark 3, 6 or 12 months down the road, and compare it with the provider running EPYC, who might have a great benchmark today, but not as much in 6 months when the node is 100% full of customers. I obviously mean the LE providers here. Not only the CPU, but we will also likely compare better on NVMe performance, network performance (transits/peerings) and datacenters. This excludes large companies (for example Hetzner) who can burn as much cash as they want and have lower operating costs since they own their DCs.
When the time comes, we will upgrade everyone to EPYC/Gold, simply be moving over the drives to the new servers. It's ~15 minutes of downtime per node. But for the moment, we deliver as good, or better performance than most LE providers running those CPUs.
Old model CPU is unbelievable cheap if you do a bulk buy. I used to work in Arrow. This is probably most LET providers' magic.
It is a good idea to start upgrading the cpu at the moment. In the foreseeable future, there is probably no cpu that can match this generation of ryzen in performance/money.
But the new hardware will put huge pressure on HH financially. Stable cash flow, which is more important than anything for a company.
HH received good amounts of hot money duing black friday, there is 100% sure they will put it into new investment(new location).
My guess is that HH may upgrade the hardware for some higher value customers, but unlikely to happen in promotional deals.
Edit: HostHatch's response faster than me.
So do I, maybe trust is gained for next year's even more fire bf offers.
@hosthatch, your VPS are very premm..
https://browser.geekbench.com/v5/cpu/search?utf8=✓&q=digitalocean is an example of what I am trying to explain above.
It will be easier to understand my above essay if you compare performance to performance in numbers, rather than CPU models.
That's fair point. Even AWS and Google can be down sometime. The best bet is to have backup somewhere else, to put a watchdog/monitoring outside DC of the main provider, and to prepare some script to do fail safe if the main route is down. Like: Auto fire up a temporary on demand node (minute/hourly billing), point domain/route to the temporary node, and redirect back when the main route is restored. All these action items can be done via API, so we can sleep well
Same here... let me know when you figured something out.
Btw: I opened a ticket to enable IPv6 for my bundle and got a reply 8 minutes after. I heard people like statistics, so: That's not even a minute per server. :P
Finally mine main server in EU is up. The CPU is not the E5-2690 v2 but still decent, especially for that price! $54/y
Thank you!
Waiting for the last one in HK)
I actually just run a docker swarm cluster so I don't have to do any of that
Just to ensure I have enough managers online to reach consensus
I'm just wondering...we see what processors a particular provider uses, starting from this we see what kind of RAM a particular provider uses DDR3 or DDR4 but we don't understand what kinda HDD or NVMe a particular provider uses. I'm really wondering, maybe someone heard any gossips about what kinda of disks local providers use? Are they enterprise or not? Are they 7200 RMP or 5400 RMP? If someone has any info about any LET provider just let me know, I'd like to extinguish my curiosity
You can always ask the providers directly
@hosthatch can you guys give us a warning if we use too much CPU? I didn't know being able to downlaod at 50MB/s will eat too much CPU.I don't want to be banned as it's a really great product.
Install Netdata and it'll warn you if CPU usage is high for a long period of time. It sends emails like this:
You can adjust the thresholds - By default it'll raise a warning if CPU usage is above 85% for 10 minutes, and a critical alert if above 95% for 10 minutes.
Make sure you modify
/etc/aliasesto forwardroot's email to an address you actually check, since it just sends the emails to the server's localroot@address. On Debian you may also need to rundpkg-reconfigure -plow exim4-configand configure it to actually send emails rather than only deliver them locally.Good that you check it, by all means. But they won’t ban you. They’ll throttle your cpu and might get in touch with you to ask what’s going on and/or offer a higher plan.
how do i know if my CPU gets throttled?
Now I am waiting for the Hong Kong one and see how it goes... never have one before in this location
Lucky you
Would you mind reminding me where the 10 locations are?
Amsterdam, Stockholm, Los Angeles, Chicago, London, Vienna, Oslo, Warsaw, New York, and Zurich. The 3x bundle is Sydney, Milan and Madrid.
Stockholm, Vienna, Warsaw, Zurich, Milan and Madrid - rarely I saw offers here - love these exotic locations though I don't know what to do with the boxes even if I had them
simply for collection
Nice purchase!
I run a site that lets you do pings, traceroutes and DNS lookups from all around the world (https://dnstools.ws/), so the extra locations will be pretty useful for my use case! I just haven't had time to set them all up yet. At least I've scripted most of the setup using Ansible so it's quite fast to spin up new servers
Good stuff! I just ran an IP with your website. Loading speed was fast.
There are several ways to throttle speed. You see reduced performance for a while and that’s it. But they are limiting to your package worth, not less than that. If you have a 50% core package, they throttle you to that.
Invoice #69777 please double cpu throttling.
You bet
@hosthatch
Eagerly waiting for the HK deployment... home sweet home...
can give a heads-up on which part of HK it would be?
Does hosthatch have a tutorial on using a custom iso? I uploaded a windows iso, but it didn’t work. I asked the technical department for help, but he has not responded to me for five days.
https://hosthatch.com/features - you can see the datacenter
Unfortunately our ISO upload feature is not very bug free for the moment due to the large number of people uploading ISOs. We are going to make it more efficient in January.
But here are a few basic points:
More importantly:
It needs to be uploaded from a fast source. If the connection breaks, then you are going to have a lot of problems;
And to be able to achieve this the best way, you can simply install your VPS with any OS first, download the ISO to it instead, put it on a webserver, and use your own URL to upload it to the control panel. Since it will be uploading to the same node where your VPS is, the likelihood of any connection problems during the upload is very low to none.
Did you read this note in the control panel: "For Windows, please make sure you add the VirtIO drivers on to the ISO. The official ISOs from Microsoft do not include VirtIO drivers by default"?
Thanks. I obviously have missed this page.
Wow! Basically the best datacentre here. 15km away from home
good stuff! look forward to seeing it going live!