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
It must be made clear that, whether the specifications are reasonable or not is a standard that we apply in our minds/actual experience; the same is true for the providers, who have not defined a reasonable range in their disclaimer.
Therefore, I am only trying to explain this reasonableness as I see it, in other words, the standard to which I am referring. This standard can be that I use the same provider, the same region, the same product (they even have almost the same name).
Man, charge 15$ for change node is very sad, even contabo doing it for free
No, it is not. I think this is where you see things very differently.
If you buy a vps with a certain specification, you should expect performance that match those specifications. If you are very lucky the performance might sometimes be higher, but this is not something that you should expect nor can demand.
HostHatch has deliver a vps well within what can be expected. Previously you were lucky and got higher performance than expected, this time you did not.
There is no "blind box", the specifications where there when you decided to buy. You also knew that there were no refunds, and you accepted that.
It is very obvious that whatever you are trying to run demands disk performance that you can normally not expect to get from a vps. It seems like you need a dedicated server and not a shared environment. I would certainly not like to share node with you, your neighbors probably suffer noticeable from you trashing the disks.
Why you think "I should" do something?
My expectation is to define it myself, not someone telling me how to define it, right? And I also explained the ins and outs of my expectations and that I don't think there's anything wrong with my expectations.
You guys are like the gods of Olympus, dictating the rules for us mortals to follow. The standards are all in your divine hands, and we must pay and be judged for our every aspiration. Even for mortals, expressing one's own emotions will likely offend the gods of Olympus, isn't that so?
To sum it up, you acted like Karen, people told you that you acted like Karen.
Your options are outlined already (no refunds, fee to change node) and server is already in expected specs and performance.
At this point, you either accept this and move on, change node by paying fee (which people said might not make any difference), transfer it to another user in a new thread.
I would say arguing with individual posts won’t change your situation and definitely wouldn’t grant you a refund
Thanks, you're just another greek god.
As I'd mentioned before: a refund was never my intention.
Regardless if you perceive me as Karen or have an imagined grasp of my intents/motivations, I explicitly stated upfront that a refund was not my goal, I was simply presenting, without judgment, the situation I encountered and at no point did I ever say I felt any wrongdoing on HH, I even eventually conceded when someone said that he thought my behavior of creating ticket for this was the reason for the delay in HH's delivery and I have even attempted to see things from HH perspective on this matter.
this totally doesnt sound like you "accept and learnt new things"
so what do you want now
actually given my own expectations I wouldn't have paid for that server, or would have at least transferred it given the high demand.
Close the thread. I've learned enough for now.
hope a mod accedes to OP's request then
Oh no! I missed a LET drama!
[On self note: don't worry! You'll get it. You'll get the next one. Calm down.]
@i12h - I'd be very interested to know the results of any profiling you've done of your actual application on this VPS vs the other one, instead of a synthetic benchmark.
For instance, start 10, 100 and 1000 consecutive requests for a typical operation - what's the average response time for a typical page on both VPSs under each scenario? i.e. what's the actual real-world difference between these two machines?
Wrong!
You should expect what the provider promises in his offer. If there are no exact performance specifications regarding the hard disk/SSD, you can't expect anything.
The part you are curious about is an area where experts delve into. With my current ability, I am not able to explain this through actual data, but maybe I will do this work one day in the future.
In the screenshot #1, I think I have been more cautious in the wording of the title of my first ticket. I just said that the "disk I/O does not look like NVMe". If I show you the current benchmark alone, will you think it looks more like SSD's benchmark?
I have nothing to do with the technical performance of the server. But the way the provider responds to the ticket is really unacceptable. You can't get the user's attention with a half-line response.
See? You guys understand this, but we mortals really don't. I just wish I had known this earlier. Thank you, I really gained something.
Which part is unacceptable?
Again. Synthetic benchmarks matter very little in the read world.
In the 1MB tests, you are getting 3.11GB/s. Quoted drive speeds from manufacturers are always contiguous large transfers. Over 3000MB/s is definitely in the range of NVMe drives, even on a dedi. You are using a VPS, this is great performance.
Even if you workload is entirely 4KB blocks, in actual practice you will get much faster access - writes will be cached in RAM until it writes to disk, and reads of frequent data will also come out of the cache in RAM. Of course, if your application explicitly waits for an fsync then things may slow down, but applications designed around performance try to avoid fsync as much as possible - for instance, read how WAL in sqlite3 can be configured to avoid fysnc.
Mostly likely you're not going to actually be hitting the disk as much as you think, unless you're doing massive searches across poorly indexed databases. If you're doing that, there are plenty of things you can do to improve performance.
But, as everybody has been saying - the performance figures you've got are very good for a VPS. When you bought the server, there weren't any promises about performance, yet you've got a great perfoming VPS. It doesn't really matter if you already have a better one, this one is great and will perform well.
No, this is a basic part of deciding what your needs are. If you don't know how much resources your application requires, worrying about whether some other machine is faster in theory than this one is the least of your worries.
I very much doubt you're doing anything that requires this much disk throughput anyway because you're going to be CPU bound if you're doing any significant processing on it, and network bound if you're not.
If you don't know how to measure these metrics, run your application and see how it performs in normal use. If it's too slow, then next time you buy a VPS, make "faster disk" a key metric you're looking out for. But realistically, few providers are going to guarantee anything faster unless you go for a dedi, and certainly not at the price point your paying.
Again, test with your application and decide if it's sufficient for your needs. Otherwise you're just bitching about a hypothetical.
This post instead subtly serves as a publicity stunt for Hosthatch 🤣 It cleverly diverts attention from their actual shortcomings and shifts the focus to highlighting the positives of HostHatch instead.
OMG, that's why I said you guys are the Greek Gods.
This thread is one of the best marketing for HH.
Even those we are regulars in criticizing HH are saying that HH is not at fault in this case.
Ew. If someone else said this, I would agree, but not you. In this thread, I only have a strong contempt for you, and you yourself said that no more reply.
if MySQL app performance is poor, one reason could be swap space. in earlier yabs, there was 4 GB swap space with 4GB memory. When it comes to MySQL, zero swap space is recommended since performance deteriorates heavily with swap space and I even included a reference link on that topic.
if I have to guess, memory is a bigger contributing factor. But one can not blame provider for that.
I agree. You’re not going to see violent conflict here, though it’s still full of posturing, with hardly anyone actually losing anything. Guess why?
That means who is saying is more important, rather than what is being said. Interesting indeed.
I did not reply to your post. You can not stop from posting that is addressed to LET members in general.
Whether you have strong contempt for me or not, it doesn’t concern me.
You're such a kid.
One can see from posts in thread and make determination who is acting like a kid.
It says vda1, a possibly volume mounted over SAN like node? A disk can be NVMe, its over network?
If you think that you can just define your own expectations and then everyone else have to live up to them, you will be disappointed. A lot.
The keyword here is reasonable expectations. Your expectations were never reasonable. To think that a vps would give you disk performance that some dedicated servers would have trouble delivering is just not reasonable.
Impossible? No, it could happen. And it seems it have happened to you, but to expect it to happen every time is not reasonable. It is not going to happen! That's like winning a million dollar every time you buy a lottery ticket.
We are not gods and we are not dictating rules, it is just common sense. You agreed to the rules when you bought the vps, we have absolutely nothing to do with that. We did not write the rules and we did not approve them, that is between you and HostHatch. HostHatch dictated the rules and you approved them!
Just face it, your expectations were not met because they were totally unrealistic. Nobody has done anything wrong here. You got what you paid for and you are not entitled to anything more.
it's paravirtualizated disk driver (virtio) and is unrelated to whether the disks are SAN or local.