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
Could you run the top command? It should give an exact output of CPU steal percentage.
Some providers use GiB, some use GB in their conversions. It really depends on the provider, but you've gotten exactly what was advertised in terms of storage. It can sometimes be confusing. Providers that give you exactly 75 GiB instead of GB are technically giving you more than advertised.
Which plan are you using? Atleast the Disk Performance is not bad in those YABS you provided, even Geekbench score is slightly nearby.
Depends on how/if it is enforced. Which is why I suggest he tries to benchmark it at other times.
But indeed, the 2 dedicated cores do not seem to be true dedicated.
If someone writes 4 dedicated cores, it means he has to provide 4 dedicated cores even if it's on a shared environment server. If they don't have this capability, they shouldn't say it's dedicated. I took service from them just because they offered a dedicated CPU. After purchasing DISK io was very high right now it's also reduced. As my application do not need much resources so it's enough. but still i'm worried what's going on there ! So far i know hosthatch is a very old provider and it's really unexpected.
this is the latest update from my 8 core (4 dedicated) NVMe vps server:
Overselling is understandable, but the server has 2 dedicated cores, all the cores are just spiking up randomly. Seems like there is nothing dedicated in that plan, so everything is shared.
Because of 6GB Swap
75 - 6 = 69
His VPS doesn't have swap. Check https://lowendtalk.com/discussion/comment/3692447#Comment_3692447
Seems op is running some type of app that generates billion$ for every GB score/ disk seek drop...
Not sure why Yabs is considered gold standard... without providing any concrete evidence that its impacting the app performance negatively. If it is, OP should get a service with dedicated cores and not a shared service
Plus I am sure he has not tried contacting support inspite of numerous recommendations... OP will simply choose to ignore the next best thing to do...
Sigh....
Request Title change... OP ran out of juice, nerfing all providers on the way
It's 4GB RAM 2CPU 30GB NVMe 1TB bandwidth with doubled RAM, storage, and bandwidth ($120/3years). So I got 8GB RAM 2CPU 60GB NVMe and 2TB bandwidth.
There is nothing gold standard about YABS, its only about the provided resources that are not the same as they were at the day 1 of purchase.
Its not running any app or even generating a penny, guess that's why it has been nerfed
because only successful business holders would be treated well by these service providers. Atleast we could arrange and pay every penny that was required for this server plan.
Should we ask the support if they "Stealing" the cpu & resources as people suggested in here?
What resources are you talking about? The provided resources appears to be the same according to both your YABS results in your post.
From a bunch of virtual servers (30+), I have/had around 3 which where oversold.
2 Providers were happy to move them.
1 Told me to gtfo, since its a "lowend server"
Price range somewhat 0.50$ to 3$, highest is 5$.
Ask them, you may have to wait, obviously, lower prio but they will surely happy helping you.
if you need better performance and support, feel free to purchase their regular priced VPS.
however they are iirc the only provider around here where GB <> GiB is concerned.
Can you advise us which one is the "regular priced" vps, because we purchased what we expected good for us using their official site only. No promotion offer or discount.
Dis provider say you will receive a certain Disk IO/ GB score when you purchased?
Why are you so obsessed with numbers if it's idling?
Ppl here have mentioned... check for cpu steal. If it exists contact support to be moved to a new node or provide a resolution.
Stop the obsession with yabs numbers. It's again, nothing more than a sense of the system at that time. Does not reflect continuous/ actual consistent usage (unless you run 24x7)
Plus it's a shared resource so there will be some degradation.
Start using the vps and see how it's impacting your usage. From the looks I am 💯 confident you will not see any issues...
Seems like this is the only way as other people also suggested same.
At least check the CPU steal, like my man here https://lowendtalk.com/discussion/comment/3692488/#Comment_3692488
this is regular price and you are entitled to make tickets for their support!!!!!!
Where do you see ANY cpu steal in these graphs?
I thought you bought a promo plan. Yes, you should have much faster support on regular plans.
??????
What CPU spikes have to do with dedicated resources/CPU steal?
Its two completly different metrics caused by completly different things.
CPU performance is normal actually.
I had a discussion on this before I made last purchase on them, and I was clearly said expectation of EPYC should be in 800-900 range on YABS.
Intel ones are expected to be in range of 500.
(Please note I am talking about G5 marks, and not G6)
Anything beyond that you see is because the Node has not been loaded at capacity yet.
OP should expect swift exodus from hatched vps. Be sure to have backups, because black box personalities does not like when there is complaints about the service...
Yes
Also note everyone ideling yet in their htop non-standard applications are visible
”Yeah 100 % idle!!!”
Htop: Docker
"I have some VPS for like $1.20 USD / month which are not oversold in any way, and performance is great."
There is NO cpu steal whatsoever in these graphs.
And of course you dont have "$1.20 USD / month which are not oversold in any way" as it wouldnt be enough to cover electricity and transit without overselling resources at least 3x, but that wasnt my point at all. Im just asking where do you see any CPU steal in these graphs as there is no such metric in these images at all.
A lower GB5/GB6 score doesn't necessarily mean that it's oversold either, I've noticed with EPYC especially that the scores can sometimes decrease once a node is more loaded depending on the core clock speeds, hyperthreading, etc
If you mean my 2 screenshots, I'm specifically showing how you can spend less than a coffee a month to have better performance than what the OP got for the price.
Those 2 VMs have 0 CPU steal. Never said otherwise.
And this is what OP's VM has idling https://lowendtalk.com/discussion/comment/3692466#Comment_3692466
How can a core be oversold if it's dedicated?
L3 cache is shared among cores
Power limit is shared among cores
Temp limit is shared among cores
RAM throughput is shared among cores
Etc. Etc. so of course even in dedicated core VPSes there is a lot of deviation between resources as Geekbench is a system benchmark that heavily relies on RAM (thats why it fails at 1GB RAM)