Howdy, Stranger!

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


LeaseWeb Reviews
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.

LeaseWeb Reviews

UmairUmair Member

Hello,

I am looking for some LeaseWeb Dedicated server reviews. (Not VPS) More like, Dedicated server on their "Volume Network"

How many of you have them and are you using them for important production sites?? I posted my "speed tests" on a different thread here for a new server I got at their NL location. No impressive. :(

Their 1st response was that network team will look into it during business hours. I am yet to hear back from them. But right now, for me their "Volume Network" doesn't seem like a network to use for production sites.

I was planning to use the server for a few VPS (fro my own setups). But really having second thought now.

Any suggestions??

Comments

  • AnthonySmithAnthonySmith Member, Patron Provider

    From what I hear LeaseWeb is great in the Netherlands, although their hardware choices are seriously questionable.

    Outside of the Netherlands where I have had a direct experience they are literally by a mile the worst company I have ever had to deal with ever, so incompetent it is beyond a joke.

    Strangely enough, the only people to deal with my stuff outside of the Netherlands in terms of support were Dutch, so perhaps it is the same people.

    Thanked by 1Umair
  • UmairUmair Member

    Well, that's what I keep hearing. That they are ok for NL location. My server with them is indeed in NL. I got the server in promo and unlike some other user experience on drives, mine are actually just a couple hours used. So no complain for server hardware.

    It's the network that I am complaining. I just got their network team response. They said they make some adjustment and my CacheFly speed should improve. (Which it actually did). However, they also said other speeds are "within expected limits". And I don't know what to do about it.

    This is the speed test I sent them.

    [root@localhost ~]# wget x86.ca/bench.sh -O - -o /dev/null|bash
    System uptime :   7 days, 9:30,
    Download speed from CacheFly: 24.3MB/s
    Download speed from OVH, Beauharnois, Canada: 9.4MB/s
    Download speed from Atlantic.NET, Toronto, Canada: 8.9MB/s
    Download speed from ColoCrossing, Chicago, USA: 8.32MB/s
    Download speed from Choopa, Piscataway, USA: 9.18MB/s
    Download speed from Leaseweb, Manassas, VA, USA: 9.1MB/s
    Download speed from Atlanta (10Gbps pipe), USA: 4.11MB/s
    Download speed from Volume Drive, Wilkes Barre, USA: 10.3MB/s
    Download speed from Atlantic.NET, Orlando, USA: 10.0MB/s
    Download speed from Incero, Dallas, USA: 7.16MB/s
    Download speed from Atlantic.NET, San Francisco, USA: 7.82MB/s
    Download speed from QuadraNET, LA, USA: 8.32MB/s
    Download speed from OVH, Roubaix, FR: 64.6MB/s
    Download speed from Redstation, London, UK: 69.1MB/s
    Download speed from Linode, Frankfurt, DE: 11.2MB/s
    Download speed from Leaseweb, Haarlem: 101MB/s
    Download speed from Linode, Tokyo, JP: 8.80MB/s
    Download speed from Digital Ocean, Singapore: 4.57MB/s
    Download speed from iiNet, Perth, AU: 522KB/s
    
  • AnthonySmithAnthonySmith Member, Patron Provider

    With the exception of Frankfurt DE they are about what I would expect, what's the problem?

  • UmairUmair Member

    Well, that above is probably the best test I could run. You didn't notice "Perth, AU: 522KB/s" or "" Atlanta USA: 4.11MB/s ""

    That is not 1Gbps ... is it??

    Have a look at these tests

    https://ibb.co/cS0pOF

    https://ibb.co/eCZcAv

    I mean, I am not expecting 100MB/s ... But I would really hope is the very least 20-25+ most of the time.

    When they are limiting user of BW usage (i.e. 10TB limit) they should not be limiting on port speed. (Or they don't have enough bandwidth there)

  • @Umair said:
    Well, that above is probably the best test I could run. You didn't notice "Perth, AU: 522KB/s" or "" Atlanta USA: 4.11MB/s ""

    That is not 1Gbps ... is it??

    Have a look at these tests

    https://ibb.co/cS0pOF

    https://ibb.co/eCZcAv

    I mean, I am not expecting 100MB/s ... But I would really hope is the very least 20-25+ most of the time.

    When they are limiting user of BW usage (i.e. 10TB limit) they should not be limiting on port speed. (Or they don't have enough bandwidth there)

    You are using US test files, please use bench.sh (freevps.us/downloads/bench.sh) for better US speedtesting, because it varies strongly from where in the US you are downloading.
    20-25mb/s US -> NL would be already very good, expect rather 12.5mb/s.
    Also, please try:
    http://mirror.i3d.net/1000mb.bin
    This would be rather acceptable for you to test portspeed as it's close.

  • UmairUmair Member

    I know I am using US test file. That test file is from LeaseWeb's Atlanta location.

    I wanted to get an overall "better network" while server being in EU (NL in this case). If I only wanted 100Mbps within NL/France, I could have gone with Ovh/Hetzner

  • ru_tldru_tld Member, Host Rep
    edited May 2017

    @Umair said:
    Well, that above is probably the best test I could run. You didn't notice "Perth, AU: 522KB/s" or "" Atlanta USA: 4.11MB/s ""

    I think you might need firstly to read about TCP congestion control, slow-start and Bandwidth delay product

    https://en.wikipedia.org/wiki/TCP_congestion_control

    https://en.wikipedia.org/wiki/Bandwidth-delay_product

    After that you may need to fix tune your sysctl settings like receive buffer or congestion algorithm so it suits for high latency connections. Connection from Netherlands to Australia is for sure a high latency connection.

    Also you might need to increase test file size. 150 MB files seems to be not enough to test 1 Gbit links on high latency transfers.

    Thanked by 3Xei Umair vimalware
  • AnthonySmithAnthonySmith Member, Patron Provider

    Yes it is 1gbit, but you will never ever ever ever see 1 gbit between the EU and Australia, you need to understand latency is a massive factor.

    Thanked by 2sin vimalware
  • @Umair said:
    I know I am using US test file. That test file is from LeaseWeb's Atlanta location.

    I wanted to get an overall "better network" while server being in EU (NL in this case). If I only wanted 100Mbps within NL/France, I could have gone with Ovh/Hetzner

    Then you might want to go to i3d with your server? They are more expensive, but their connection to both US and Asia are really good for a NL server, my one:

    Location Provider Speed
    CDN Cachefly 111MB/s

    Atlanta, GA, US Coloat 22.2MB/s
    Dallas, TX, US Softlayer 22.9MB/s
    Seattle, WA, US Softlayer 19.9MB/s
    San Jose, CA, US Softlayer 20.8MB/s
    Washington, DC, US Softlayer 16.5MB/s

    Tokyo, Japan Linode 10.04MB/s
    Singapore Softlayer 11.9MB/s

    Rotterdam, Netherlands id3.net 112MB/s
    Haarlem, Netherlands Leaseweb 101MB/s

  • UmairUmair Member

    @ru_tld said:

    After that you may need to fix tune your sysctl settings like receive buffer or congestion algorithm so it suits for high latency connections. Connection from Netherlands to Australia is for sure a high latency connection.

    Also you might need to increase test file size. 150 MB files seems to be not enough to test 1 Gbit links on high latency transfers.

    Agreed. But even without needed any special tweaks, it should work better on default CentOS 7. And I get that large files would be better, but 100MB file should give a fair idea.

    Don't we all post output of freevps bench everywhere when asked for network performance??

    @AnthonySmith said:
    Yes it is 1gbit, but you will never ever ever ever see 1 gbit between the EU and Australia, you need to understand latency is a massive factor.

    @Anthony, I understand that. I was not expecting anything extra ordinary to AU either. But still 522KB/s is just too low for any "decent" network. That along with the fact that my download to their own US network is fluctuation a lot. Like 60% of the time, I get speed in KB/s fro US links and the rest 40% of time, its between 5-8MB/s. Would you call this a decent network ????

    Look at this result I just did

    [root@localhost ~]# wget -O /dev/null http://chi.testfiles.ubiquityservers.com/100mb.txt
    --2017-05-30 18:20:18--  http://chi.testfiles.ubiquityservers.com/100mb.txt
    Resolving chi.testfiles.ubiquityservers.com (chi.testfiles.ubiquityservers.com)... 174.34.172.51
    Connecting to chi.testfiles.ubiquityservers.com (chi.testfiles.ubiquityservers.com)|174.34.172.51|:80... connected.
    HTTP request sent, awaiting response... 200 OK
    Length: 104857600 (100M) [text/plain]
    Saving to: ‘/dev/null’
    
    100%[=======================================>] 104,857,600 96.1KB/s   in 17m 41s
    
    2017-05-30 18:37:59 (96.5 KB/s) - ‘/dev/null’ saved [104857600/104857600]
    

    @bluesega said:

    Then you might want to go to i3d with your server? They are more expensive, but their connection to both US and Asia are really good for a NL server, my one:

    Yeah well I am seriously considering alternatives. The price I got with LeaseWeb was awesome. (for the server I have). I was not expecting extra ordinary network performance ... but the test I have done so far are way too poor for me to use this server for any high traffic site.

    @AnthonySmith , @bluesega , @ru_tld
    One of the reason I am complaining is ... Compare the above speed test with the following from $5/m vps via Vultr

    [root@testvm ~]#  wget x86.ca/bench.sh -O - -o /dev/null|bash
    CPU model :  Virtual CPU a7769a6388d5
    Number of cores : 1
    CPU frequency :  2394.454 MHz
    Total amount of ram : 992 MB
    Total amount of swap :  MB
    System uptime :   27 min,
    Download speed from CacheFly: 87.3MB/s
    Download speed from OVH, Beauharnois, Canada: 12.5MB/s
    Download speed from Atlantic.NET, Toronto, Canada: 11.20MB/s
    Download speed from ColoCrossing, Chicago, USA: 12.66MB/s
    Download speed from Choopa, Piscataway, USA: 14.1MB/s
    Download speed from Leaseweb, Manassas, VA, USA: 23.4MB/s
    Download speed from Atlanta (10Gbps pipe), USA: 25.7MB/s
    Download speed from Volume Drive, Wilkes Barre, USA: 23.1MB/s
    Download speed from Atlantic.NET, Orlando, USA: 9.71MB/s
    Download speed from Incero, Dallas, USA: 13.0MB/s
    Download speed from Atlantic.NET, San Francisco, USA: 13.3MB/s
    Download speed from QuadraNET, LA, USA: 7.4MB/s
    Download speed from OVH, Roubaix, FR: 126MB/s
    Download speed from Redstation, London, UK: 74.5MB/s
    Download speed from Linode, Frankfurt, DE: 17.2MB/s
    Download speed from Leaseweb, Haarlem: 287MB/s
    Download speed from Linode, Tokyo, JP: 9.22MB/s
    Download speed from Digital Ocean, Singapore: 12.5MB/s
    Download speed from iiNet, Perth, AU: 4.62MB/s
    
  • ru_tldru_tld Member, Host Rep

    I don't think at all that your benchmarking is somehow relevant to real network performance. Indeed downloading small files from may be busy location can indicate some network issues but even with your results I don't see any.

    but the test I have done so far are way too poor for me to use this server for any high traffic site.

    Do you suppose that high traffic sites are relying on single tcp connection performance on high latency routes?

    I think you are over-engineering your project.
    But If you really want to follow this way, you can try to compare 1x thread/core performance of 2650v5 cpu with performance of 1650V4 cpu...There will be a surprise.

  • UmairUmair Member

    @ru_tld said:
    I don't think at all that your benchmarking is somehow relevant to real network performance. Indeed downloading small files from may be busy location can indicate some network issues but even with your results I don't see any.

    No, that does not really show "real network performance" but it DOES give you an estimate. (i rough idea of how things going to be). That's what I am doing here as well.

    Do you suppose that high traffic sites are relying on single tcp connection performance on high latency routes?

    I think you are over-engineering your project.

    May be I am... But that's what I do before moving any of my client's data to a new Host I have not tried before. I have not yet moved any live setups to them since I am still "testing them".

    When my connection speed would very between 500Kbps to 8Mbps (while mostly being around 500-900Kbps) on a 1Gbps connection, I would be concerned.

    Have you seen the result of download I posted above?? "(96.5 KB/s)" seems normal to you from Chicago to NL??

    And yeah, I know about latency / single threaded download / etc etc etc.

  • AnthonySmithAnthonySmith Member, Patron Provider

    Fair enough, honestly it seems reasonable to me for real world use.

  • XeiXei Member

    @bluesega said:

    @Umair said:
    I know I am using US test file. That test file is from LeaseWeb's Atlanta location.

    I wanted to get an overall "better network" while server being in EU (NL in this case). If I only wanted 100Mbps within NL/France, I could have gone with Ovh/Hetzner

    Then you might want to go to i3d with your server? They are more expensive, but their connection to both US and Asia are really good for a NL server, my one:

    Location Provider Speed
    CDN Cachefly 111MB/s

    Atlanta, GA, US Coloat 22.2MB/s
    Dallas, TX, US Softlayer 22.9MB/s
    Seattle, WA, US Softlayer 19.9MB/s
    San Jose, CA, US Softlayer 20.8MB/s
    Washington, DC, US Softlayer 16.5MB/s

    Tokyo, Japan Linode 10.04MB/s
    Singapore Softlayer 11.9MB/s

    Rotterdam, Netherlands id3.net 112MB/s
    Haarlem, Netherlands Leaseweb 101MB/s

    bluesega what provider are you using? Those speeds are better then I expected for i3d. My friend picked up a NL VPS from popular LET provider in Smart DC also. Stress testing over a week or so speeds were nowhere close to your post and varied a lot depending on time of day. How consistent would you say speeds from the US are?

  • bluesegabluesega Member
    edited June 2017

    @Xei said:
    bluesega what provider are you using? Those speeds are better then I expected for i3d. My friend picked up a NL VPS from popular LET provider in Smart DC also. Stress testing over a week or so speeds were nowhere close to your post and varied a lot depending on time of day. How consistent would you say speeds from the US are?

    It's @Verelox

    Don't know much about usual US peering though, might be at a lucky time, but the **least ** I've got yet during US speedtests was about ~50mbit/s.
    Don't know if they apply some magic onto i3d's servers in order to speed them up some more, but speeds are definitely at least acceptable for EU->US. (EU->EU is great, EU->Asia is better than most other providers in Europe)

    Speedtest from a minute ago:

    Location Provider Speed
    CDN Cachefly 111MB/s

    >

    Atlanta, GA, US Coloat 9.66MB/s
    Dallas, TX, US Softlayer 21.9MB/s
    Seattle, WA, US Softlayer 17.9MB/s
    San Jose, CA, US Softlayer 17.7MB/s
    Washington, DC, US Softlayer 6.74MB/s

    >

    Tokyo, Japan Linode 9.10MB/s
    Singapore Softlayer 11.1MB/s

    >

    Rotterdam, Netherlands id3.net 112MB/s
    Haarlem, Netherlands Leaseweb 95.4MB/s

    It varies for the US, but all in all it's nice.

    Thanked by 1Umair
  • UmairUmair Member

    @bluesega

    That indeed looks nice speed. PS. You may want to use Code tag to format the results output.

    On a side note, I just learned that LeaseWeb counts both incoming + outgoing bandwidth towards their 10TB limits. (Looks like I didn't read things clearly 1st time)

    So I am looking at additional 25 Euro for monthly bw usage :( :(

Sign In or Register to comment.