Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


1 Year with Vultr: My review
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.

1 Year with Vultr: My review

I'm currently moving away from Vultr (I'll explain why further down) and decided to write a review as a farewell.

After 1 year with Vultr I've had a total of 2 problems that were out of my control. Both problems were high latency and packet loss, but they only lasted for about an hour each. So with around 7 servers for 12 months, I've had 2 hours of interruptions. Not bad at all. Their support was average at best with these 2 problems, it took over an hour for them to respond and it was just the usual response of "Sorry but it's not our fault" (paraphrase).

I use some of my Vultr instances as game servers, and I've seen people playing on the server with 3ms ping which clearly proves their networks can quickly handle packets and I've never seen latency spikes aside from the 2 mentioned above. Although the game server doesn't use much IO, I always found the SSD to be fast and reliable.

I've had a total of 6-7 DDoS attacks on these Vultr instances. They don't offer DDoS protection, and I wasn't using an external service for it. When getting hit with attacks of around 10gbps, I was simply null routed and Vultr never complained. I was expecting some kind of warning letting me know I'd be suspended if the attacks continued, but that never happened. Whenever a server has been null routed, their support is fairly useless. If you ask them how long it'll be until your connection is working again, they won't respond until after the null route is lifted and they'll just say something along the lines of "removed null route" and close the ticket (hours after it got opened).

I've also used Vultr instances for very heavy CPU/IO tasks in the past. Even after hours of using 100% IO and high CPU, I never got any complaints or messages. They genuinely seem happy to let you use as much resources as you want which means you don't have to worry too much about your usage. I've not once felt like they're overselling their servers, and it really does feel like the resources you get are dedicated to you.

The reason why I'm now moving away from Vultr is because they've suddenly dropped support for their 3.5GHz processors (and are now only offering 2.4GHz) and they've increased the price by 20%. I'm not going to pay more money for slower servers.

In terms of their reliability, I have to give them a 5/5. Their servers are simply rock solid. If I wasn't manually rebooting these servers on occasion, they'd have 12 months uptime. Considering the worst problem I ever had with them was an hour of high network latency, there's really no complaints in regards to their hardware.

I'd give their support a 2/5. It feels like robots are answering all the tickets, but I bet if you had a problem they'd do whatever they could to help even if it does take them close to an hour to respond.

The fact that they suddenly upped the price and lowered the server specs overnight is a little off-putting. However, if you're looking for a reliable server with a 2.5GHz processor, Vultr will probably serve you well.

Thanked by 1josephb

Comments

  • J1021J1021 Member

    inb4 @kcaj

    Thanked by 2Amitz ATHK
  • trvztrvz Member
    edited May 2015

    But you already had the 3.4-3.6 GHz servers, they wouldn't get downgraded (except in London) and they stay at the old price. Why not keep them?

    What other provider is out there with 3.4+ GHz servers, that is not oversold and which you can use with constant full load?

    Thanked by 2lucast J1021
  • TarZZ92TarZZ92 Member

    trvz said: What other provider is out there with 3.4+ GHz servers, that is not oversold and which you can use with constant full load?

    vultr am sure oversell. and am sure they don't put 4 VPS per node, since most of these E3's are 4 core 8 threads. so technically they are overselling and treating threads as cores

  • TarZZ92TarZZ92 Member
    edited May 2015

    Jack said: What was you doing that attracted DDoS?

    he said game server.. we (game server owners) all have to put up with this, and they usually come from competitors

  • ChanChan Member

    Mind telling us where you've moved to?

  • @trvz said:
    But you already had the 3.4-3.6 GHz servers, they wouldn't get downgraded (except in London) and they stay at the old price. Why not keep them?

    What other provider is out there with 3.4+ GHz servers, that is not oversold and which you can use with constant full load?

    Not Vultr

    Thanked by 1W3HostingServices
  • DO also downgraded their node CPUs and still nobody complains about it:

    CPU model :  Intel(R) Xeon(R) CPU E5-2650L v3 @ 1.80GHz
    Number of cores : 1
    CPU frequency :  1797.917 MHz
    Total amount of ram : 490 MB
    Total amount of swap : 0 MB
    System uptime :   1 day, 20:07,       
    Download speed from CacheFly: 62.4MB/s 
    Download speed from Coloat, Atlanta GA: 12.3MB/s 
    Download speed from Softlayer, Dallas, TX: 15.4MB/s 
    Download speed from Linode, Tokyo, JP: 5.66MB/s 
    Download speed from i3d.net, Rotterdam, NL: 32.4MB/s
    Download speed from Leaseweb, Haarlem, NL: 128MB/s 
    Download speed from Softlayer, Singapore: 7.23MB/s 
    Download speed from Softlayer, Seattle, WA: 12.6MB/s 
    Download speed from Softlayer, San Jose, CA: 11.7MB/s 
    Download speed from Softlayer, Washington, DC: 4.63MB/s 
    I/O speed :  381 MB/s
  • TarZZ92TarZZ92 Member

    Vultr is totally different. u can also ur own os on vultr for example

  • yomeroyomero Member

    alexvolk said: DO also downgraded their node CPUs

    On what location? That seems like NL.
    I use the usa locations and I get 2-2.3Ghz cpus IIRC

  • @yomero said:
    I use the usa locations and I get 2-2.3Ghz cpus IIRC

    Germany.

  • J1021J1021 Member

    TarZZ92 said: Vultr is totally different. u can also ur own os on vultr for example

    Lack of ISO support at DO allows yourself to see them in a better light?

  • TarZZ92TarZZ92 Member

    kcaj said: Lack of ISO support at DO allows yourself to see them in a better light?

    it's not just ISO it's the kernel too.

  • alexvolk said: DO also downgraded their node CPUs and still nobody complains about it:

    DO German 1.80GHz UnixBench score better than NL 2,4Ghz

  • jarjar Patron Provider, Top Host, Veteran
    edited May 2015

    @dragon1993 said:
    DO German 1.80GHz UnixBench score better than NL 2,4Ghz

    Good eye. Clock speed is not the only way to judge a CPU. :)

  • @trvz said:
    But you already had the 3.4-3.6 GHz servers, they wouldn't get downgraded (except in London) and they stay at the old price. Why not keep them?

    For $1/month less, I can get a processor with the same speed and an extra 2 cores + DDoS protection from RamNode. I'm also worried that one day Vultr will decide to move my server onto slower hardware, I asked if they would do that on their forum and got no response, so I'm assuming they will.

    Based on how often I create/destroy or move servers, it's quite annoying to have to log in to multiple providers for something that Vultr used to be able to do single handedly.

    I have nothing against Vultr, I just can't use them for any of my new servers, and I'd get better value for money elsewhere, so I see no reason to stick around.

    @Jack said:
    Surely after the 2nd null-route though you'd be like p'd off and look into getting 3rd party mitigation in place?

    Not particularly viable for game servers. Most of them add 20-30ms latency onto the connection depending on the provider/location, then there's the regular latency spikes. I only managed to find about 3 providers under $100/month that could cover game servers. I've always found things to be more reliable when you find a host with built-in DDoS protection, even if they're using a 3rd party service.

    Anyway, it was around 6 attacks in the space of a year across multiple servers so it was barely a disruption. Annoying at the time, but not a big problem when you see the big picture.

    @Chan said:
    Mind telling us where you've moved to?

    RamNode is looking very promising. I moved one server over to them and after 3-4 weeks I've not had a single problem.

    Thanked by 1Nick_A
  • J1021J1021 Member

    Josephhewitt said: For $1/month less, I can get a processor with the same speed and an extra 2 cores + DDoS protection from RamNode.

    Whether you have 1, 2 or 4 cores at RamNode doesn't make much difference.

    As in all virtual server environments, CPU cores and disk I/O are shared among multiple VPSs. If your VPS is consistently maxing out one full core or more and impacting the performance of other client servers sharing the same host node, we may ask you to reduce your usage. Load may be generated by CPU usage and/or I/O. Disruptive load may result in a reboot, shutdown, and/or suspension of the VPS regardless of the time lapse involved.

    In any case you're only permitted to use 1 core for sustained periods of time.

  • @kcaj said:
    In any case you're only permitted to use 1 core for sustained periods of time.

    I know. Me using 100% IO for hours was an edge case, I have dedicated servers for that now. I just thought I'd mention that in the review in case people were considering Vultr for high IO/CPU applications. It was something I was testing on a $5/month Vultr server rather than a $50/month dedicated server.

    I've had a small (128MB) server with RamNode for around 6-7 months without issue, and a 4GB server with them for a month without signs of problems.

    While I realise RamNode is more 'buget' than Vultr, their performance so far seems close to equal. The fact that RamNode offers much faster cores than Vultr now makes them seem a lot more attractive

  • AnthonySmithAnthonySmith Member, Patron Provider

    I don't think the clock speed alone is a valid judgement at all, and frankly in 2015 on a VPS forum I am pretty stunned that seems to be a deciding factor.

    Would you rather have access to an X3470 @ 2.93Ghz from 2009 or a Pentium 4 @ 3GHZ from 2004 or an E5-2667 which is slower than both of them at only 2.9 GHz?

    Also you need to keep in mind they have WAY more than 8 VPS per node (8 logical cores on an E3), so while the clock speed may well be faster it's cycles are probably contended about 400% more than on a dual E5 based server.

    Give me faster Ram, better IOPS and less contended cores over the raw clock speed of a single logical core any day.

    Thanked by 3jar Lee alexvolk
  • @TarZZ92 said:

    DO allows custom kernel. Just send a ticket.

    Thanked by 1vimalware
  • jarjar Patron Provider, Top Host, Veteran

    @TinyTunnel_Tom said:
    DO allows custom kernel. Just send a ticket.

    With Ubuntu 15.04, Debian 8, CoreOS, and FreeBSD the VMs control their own kernels instead of loading them externally. Sorry, not trying to hijack thread, just let you know :)

  • @Jar said:
    With Ubuntu 15.04, Debian 8, CoreOS, and FreeBSD the VMs control their own kernels instead of loading them externally. Sorry, not trying to hijack thread, just let you know :)

    I asked about Debian 7 and CentOS 6 thats why ;)

    Thanked by 1jar
  • TarZZ92TarZZ92 Member

    TinyTunnel_Tom said: DO allows custom kernel. Just send a ticket.

    with vultr no need for that :)

Sign In or Register to comment.