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
thanks @cociu
Uhhr nice .. is it Serbia ?
yes , after this focusing in open our first africa location.
I am on prefix 45, hope to get it resolved soon.
we estiate 3 days , and a short downtime until the node is rebooting (after the aditional ip is added to all clients we will reboot the node to take effect the new settings ) so depend of wich node you are will be more rappid or up to 3 days. Right now for exemple the first node called stor17 is done so will be rebooted in some minutes.
Sounds awesome
If u ever consider Moldova that might be something for u
https://trabia.com/colocation.42-0.html
we have moldova like 5 years ago for some "in house" clients ... maybe i will send some servers there and make it publick ... but not necesarry.
that sounds amazing good luck
Has the CPU performance issue been resolved, and is there any discount for nvme vps now?
Mon Jan 4 19:27:37 UTC 2021 Basic System Information: --------------------------------- Processor : Intel(R) Xeon(R) CPU E5-2680 v4 @ 2.40GHz CPU cores : 4 @ 2397.222 MHz AES-NI : ✔ Enabled VM-x/AMD-V : ❌ Disabled RAM : 3.9 GiB Swap : 0.0 KiB Disk : 246.0 GiB Running GB5 benchmark test... *cue elevator music* Geekbench 5 Benchmark Test: --------------------------------- Test | Value | Single Core | 245 Multi Core | 258 Full Test | https://browser.geekbench.com/v5/cpu/5704187I got AES-NI enabled, but openVPN performance is pretty bad. Iperf directly on VPS is reporting 60+ MBit/sec, iperf locally without VPN is also 40+ Mbit/sec, Iperf over OpenVPN is reporting under 6 MBit/sec.
Though CPU usage on the VPN never goes above 70%. Not sure why OpenVPN sucks. Any suggestion?
Are you confident in your openvpn setup?
I haven't run many tests, but my openVPN doesn't seem to have any issues.
He brings great shame to his family.
Similar configuration works fine on other VPSes. This is what I've,
Could it because I ordered the cheapest configuration of 256GB?
What? A lower multi core score than single core on a multicore server would be the exact symptom of overselling. A sign CPU isn't oversubscribed when the multicore score is roughly a multiple of single core score.
Also, those are really, really low scores to begin with. It's definitely under sized for the load.
Yes, don't use openvpn. Use wireguard.
I am not thinking about overselling because every time I get almost the same benchmark results. If server is oversold, results definitely will not be the same everytime.
its a new node with verry limited customers there , more precise 38 customers in this node , what i suspect finnaly is a missettings what hp have in bios and we will try to modif this today . Definetly is not overselling. in proxmox settings we have modif all what we know and no results. More in a new node is give the same results.
May I know the node specs? @cociu
2xE5-2680V4
128 gb ddr4
4x2tb nvme
4x12tb sata
port speed?
??
2 x 14 Core = 28 Core ( 56 vCore ) / 128 GB DDR4 - have
38 vps x 4 vCore/4 GB RAM = 152 vCore / 152 GB DDR4 - sell
Man.....I guess my two VPSes are in the same node, one is 1C/2G, the other is 2C/2G. So, I believe RAM is not overselling. As for CPU cores, they are differently controlled. If you had ever played with Proxmox, you would know that.
The current problem of cpu performance is definitely not caused by overselling. This CPU won't have such low single core performance.
no worry about them , all is verry smart its seems ...
@telimp i can sell even 16 cores or including more but where you have see when i have mentionated the quantity of cosres alocated/vm ?
@cociu
When a node is done with the 45.14.x.x change the client panel will reflect the changes, correct?
simple math....
If you talk about vCore than is important how you share the 56 vCore, you can have right
"Core" and "vCore" is commonly used interchangeably. 28 hw cores are 56 hw threads which are commonly understood as "dedicated cores", while more often than not only fractions of a hw thread are sold as "vCores" (shared). So, your "hunt" is meaningless.
I've tried to advocate for providers providing more concrete specs than "vCore", in particular how much of a hw thread a "vCore" actually is, but it seems that most didn't care and didn't support me.
That said @cociu at least provided a basic node spec, which is much better than what most other providers offer.
In the vps industry, vcore is always referred to as core. Sometimes you get real core and x2 vcore in virtual dedicated servers depending on the definition of the provider.
With vps, it is always fair-share cpu unless otherwise mentioned, so cpu is always 'oversold'. But as long as the load on the server is managed well, it's acceptable. Also, out of the 38 clients, some may have 1 core or 2 cores. Not all of them have 4 cores. Yes, it is overselling but not in a bad way.
Looking at the geekbench, I think cociu made it so you get half the cpu cycles, so 50% of a vcore. But maybe I am wrong. Maybe there's some config issue so the cpu is slow.
once you note downtime you can take a look to the pannel for the new ips. we need to restart each node after change
updte : we are missing arrownd 180 ips to change (5 min for each ip is tard) so maybe we are done today.
Go go Marius! I received my new IP and uptime was solid after that! Nice.