All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Terrible disk speed, or expected for oversold VPS?
I signed up with woog to get a small VPS on their LEB deal. My hopes was to get a small wordpress site set up for a local tree service guy (swapping labor).
Anyway, after I purchased the server, I waited about a week to start working on it. Most of the reviews at the time were positive, and I didn't need lots of horsepower for such a small site...so I didn't test within the limited refund window.
Shortly after getting started, I realized that I was having horrible 30s+ TTFB issues and other seemingly latency related problems with SSH, etc. My ping response and cachefly speeds were great, however (108 MB/s).
I ran a disk i/o test and came up with some perceptibly low numbers. My first test was around 6-9MB/s, so I opened a ticket with support inquiring about the issue. Support was quick to respond and "optimize the node". I tested again and got up in the 40MB/s range. Ok, not blazing fast..but doable, or so I thought.
After a week, I went to do some more work from the frontend and started noticing the issue crop up again. So, I opened another ticket after getting frustrated with the same problem. Once again, support was quick to respond. However, this time they claim no issues from there end. When I went to check, speeds were reasonable again.
A day later, I am logging in to check and get 1.6MB/s. I've had this server for two weeks now, and am unsure of what to do. Is this a sign of a highly oversold VPS? I've not had many, and this is the first time that I've needed to open tickets for an issue like this. Is anyone else getting slow disk speeds?
Comments
Are you running those speed tests at the same timestampt as your previous tests/days?
No, this has been on different days/times. Although the times I have tried are usually between the hours of 1-4am.
Me too, I have vps on them, when I backup my files, its so slow. Im backing up worth 200mb only.
What's the pricing you're paying for this and what location? I'd definitely look towards another provider if I were you... You've reported the issue and they have said they have fixed it a week later you're getting the same stuff.
It's clearly a recurring problem and they're likely just suspending users who are using a bit too much IO.
LET has some fantastic choices of providers out there and I'm sure you can find something very similar to what you want/need.
My guess would be that everyone on that node could be running backups during those times, it is common to run backups during midnight/aftermidnight try running those tests on different times other than 1 am - 4 am to get an idea of what might be going on.
I just finished scp my webfiles to a different provider (alpharacks) and it took over 30min for 300MB. Ran another test just now and got 12MB/s.
If it was network related, I could graph it. I'm just not sure if there are any utilities to check and graph the disk issue. All I care about are reasonable response times to the website during living people hours. As a night shift worker, it is impossible to work on this site. I set up pingdom to check site availability, but do you guys know of something I can do to provide evidence to support about this? Rivermigue, I've considered the backup times being the source of the problem..but I'm a night shift worker, so I need to find a way to make due.
Just ran a test on my similar spec alpharacks vps and got 46.3 MB/s.
To respond to ithinkufailed: it was their $14/yr low end special. The server is located in Los Angeles.
I've had good experiences with LEB, using crissic and weloveservers. This is the first time I have ever bothered to check benchmarks, since it is painfully noticeable.
$14/yr is a pretty low cost plan but I'd still say 12MB/s is unreasonably slow. I'd ask for a refund and migrate elsewhere honestly. If it's for a business, even if small, you need some reliability and 12MB/s IO isn't reliable for anything. This might mean paying a bit more or picking a cheaper location (ArubaCloud comes to mind)
@IThinkUFailed I agree with you regarding reliability. I believe I will politely ask for a refund. Thanks.
I wish you luck with that. At the end of the day, it's $14 and you come off more prepared in future for testing a VPS if it suits your needs or not
I'm prepared to chalk it up to a learning lesson, either way. =P
It may be worth a try: @wootng (I think that the guy's name is Jason.)
This still seems a somewhat slow speed, 50MB/s is considered ok but you need a host that will enforce IO usage limits to make sure you have the IO when you need it, Most hosters will allow X amount average over say 5 minutes however on a side note sometimes this can't be helped if someone is abusing but the company should also be monitoring.
My advice here is to find a host that enforces there io limits, maybe try and find a host that uses SSD's as this mean the io usage is somewhat less likely to be abused as there is more in general to use.
Thanks @angstrom for the tip. The company is kind and responsive, so I hate to leave a negative review from my experience. Perhaps the $14/yr is more geared toward networked based services such as irc, vpn, etc?
Also, thanks @Bopie for the insight and suggestions. I have been looking at some low end SSD options. I'm not sure if providers' IO limit enforcement policies are publicly available (in my limited experience). Do SSD cached servers generally offer similar performance to pure SSD servers?
@sv3nster
the IO enforcement should be public in maybe an AUP or TOS not to shamelessly plug my own company, however, our AUP makes it very clear on what you can and can't do on our VPS this includes IO limits etc.
As for SSD Cached, you won't see the same performance as pure SSD but it well be in between both HDD and SSD I would imagine I couldn't tell you for sure as I have not tested it.
Hmmm. There's nothing in their policy docs regarding IO specifics, but they do cover standard forbidden CPU hoarding activities. Nothing wrong with plugging your company, whilst assisting and enlightening the public at the same time. I suppose if more providers gave explicit technical details such as you do, we would all be a little more informed at the end of the day. I will consider this in the future. Thanks for @Bopie helping.