Howdy, Stranger!

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


Building a Serverbear alternative
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.

Building a Serverbear alternative

AntonAnton Member
edited June 2016 in General

Hey guys,

I'm working on an automated benchmark tool for Linux servers and have just finished the MVP, so I figured it would be a great time to get some feedback.

Basically, I want to eventually build a modern better version of Serverbear since it's been pretty much abandoned for the last two years.

At this point I only have two pages:

In the next couple months I plan to finish up comparison pages that would allow users to compare aggregated benchmark results between different VPS offers, run lots of benchmarks to collect more data (as I don't expect this thing to become popular any time soon), and add a few other benchmarks into the benchmarking script as it currently only includes dd, fio, unixbench and a bandwidth benchmark via speedtest.net.

The benchmark tool comes as a Python script and here is what it does:

  1. Checks for dependencies and installs them if needed. Depending on the platform dependencies are either make, automake, gcc, gcc-c++, kernel-devel, libaio-devel (for yum) or build-essential,libaio-dev (for apt)
  2. Gets the location of the server from API using curl
  3. Collects some server specs using Linux tools like df and output of /proc/cpuinfo / proc/meminfo
  4. Runs benchmarking tools (fio, dd, unixbench and a custom bandwidth test)
  5. Uploads the results and returns a link to the report page

I would appreciate all kinds of feedback, especially things that you would like to see implemented such as benchmarks or website features.

«1

Comments

  • Running one right now. Looks interesting so far

  • Finished the bench. No link returned though.

    Thanked by 1EricB
  • Mark_RMark_R Member

    that result page looks really awesome man, nice design!! ill try it out.

    Thanked by 1Anton
  • Mark_RMark_R Member
    edited June 2016

    i ran into an little error while running the commandline provided from serverscope.io.

    root@w24:~# curl -O "https://raw.githubusercontent.com/anton-ko/serverscope-benchmark/master/serverscope.py" && python serverscope.py -e "[email protected]" -p "VPSPro basic + ddos protected IP@SeFlow"
      % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                     Dload  Upload   Total   Spent    Left  Speed
    100 16971  100 16971    0     0  27750      0 --:--:-- --:--:-- --:--:-- 28570
    Need to install libaio-dev
    Installing libaio-dev
    Traceback (most recent call last):
      File "serverscope.py", line 274, in 
        subprocess.call(['sudo', 'apt-get', 'update'])
      File "/usr/lib/python2.7/subprocess.py", line 493, in call
        return Popen(*popenargs, **kwargs).wait()
      File "/usr/lib/python2.7/subprocess.py", line 679, in __init__
        errread, errwrite)
      File "/usr/lib/python2.7/subprocess.py", line 1259, in _execute_child
        raise child_exception
    OSError: [Errno 2] No such file or directory
    root@w24:~#
    

    if you need more information let me know.

    Thanked by 1Anton
  • Most of the people use servers as web server or apps based on php / apache /nginx , when you say benchmarks, may be you should include such parameters also like how much the vps can serve.

  • AntonAnton Member

    @Saragoldfarb said:
    Finished the bench. No link returned though.

    Ugh, I'm sorry about that. Could you share provider / plan and OS? I would really like to troubleshoot that. Thank you!

    @Mark_R said:
    i ran into an little error while running the commandline provided from serverscope.io.
    ....
    if you need more information let me know.

    Hey Mark, thanks for the report. Could you also tell me what OS / version you're running?
    Thanks

  • AntonAnton Member

    @slicebox said:
    Most of the people use servers as web server or apps based on php / apache /nginx , when you say benchmarks, may be you should include such parameters also like how much the vps can serve.

    I agree and I've been trying to come up with a way to implement something like that without requiring too many dependencies. My best idea so far was to put the app into a docker container but that wouldn't work on OpenVZ servers.

    I would really appreciate a suggestion on this. Or you would be okay with script installing LAMP? To me that just seems a bit too intrusive.

  • All done! Submitting the results...
    {"status":"500","error":"Internal Server Error"}  
    
    Thanked by 2pbgben Anton
  • installing LAMP would be nasty, what if the vps already has LAMP, haven't thought deep into this, it is bit complicated. unless everything can be extracted to a folder , run from there, then delete after completion, (stopping all usual services before running test)

  • The network part of the result page is a bit confusing. The raw output below says 302 Mbit/s, while above it says 302 MB/s.

    Thanked by 1Anton
  • pbgbenpbgben Member, Host Rep

    @Strikerr said:
    All done! Submitting the results...
    {"status":"500","error":"Internal Server Error"}

    This happened to me too, on the OVH network.

  • LiteServerLiteServer Member, Patron Provider

    Looks good! I'll run a test shortly and let you know my feedback.
    Do you have plans to make a page with a "top 100" to compare CPU / RAM / IO performance - basically the same as what Serverbear has? I think such page will definitely have a value as people might want to compare performance of multiple hosting companies.

  • Mark_RMark_R Member

    @Anton said:

    Hey Mark, thanks for the report. Could you also tell me what OS / version you're running?
    Thanks

    Debian 7.5 x64

    Thanked by 1Anton
  • TomTom Member
    All done! Submitting the results...
    {"status":"500","error":"Internal Server Error"}
    
    Thanked by 1inthecloudblog
  • AntonAnton Member

    Ugh. Okay. Thanks for the reports guys.
    I'm looking into that error 500.

    @salakis said:
    The network part of the result page is a bit confusing. The raw output below says 302 Mbit/s, while above it says 302 MB/s.

    Great catch! Will fix that today as well.

  • shovenoseshovenose Member, Host Rep

    Thank you... Serverbear is broken.

    Thanked by 1Anton
  • Don't really know that much when it comes to servers and stuff, but love the initiative! I'll make sure to bookmark this for future reference.

    Just one question though, will this a free tool or will this be a paid service? Thanks a lot!

  • cassacassa Member

    Works pretty well: https://serverscope.io/trials/nAr

    The disk read test didn't go right though.

  • hawchawc Moderator, LIR

    The links in both of my emails didnt work.

  • AntonAnton Member
    edited June 2016

    @rgenzon said:
    Just one question though, will this a free tool or will this be a paid service? Thanks a lot!

    Will always be free for users. Glad you liked it!

    @hawc said:
    The links in both of my emails didnt work.

    Thanks for reporting back. Would you mind sending me your link? I will make sure it's fixed.

    Thanked by 1rgenzon
  • cassacassa Member

    @hawc said:
    The links in both of my emails didnt work.

    They're correct for me

  • SilvengaSilvenga Member
    edited June 2016

    It might be a good idea to have consistent speed test servers so that we can compare apples to apples across different locations - it would also prevent issues when geo-location fails.

    Also it would be nice to include all the human readable stats at the top of the page, rather then strewn throughout.

    Thanked by 2Anton cassa
  • AntonAnton Member

    @cassa said:
    They're correct for me

    There was an issue with benchmark on @hawc's server. Has been fixed.

    @Silvenga said:
    It might be a good idea to have consistent speed test servers so that we can compare apples to apples across different locations

    Thanks for raising this issue. I like this! My only concern is that people usually host their servers close to their users, so my assumption is that they care primarily about bandwidth between the server and major providers around it. For example, if I build a website for Canadians I would care only about loading speed for users in North America.

    I don't have a solution for this yet. My thinking behind this was that if I take 10 closest servers (and skip the closest one) then I benchmark the connection speed between the host and the outside world. I'm not 100% satisfied with this as major cities would often have 10+ speedtest servers located very close so results are skewed. I'm thinking that I might change it so it skips all the servers that are too close (let's say 10 or 20 km). Would that be a good idea?

    Or may be I should pick a few sets of servers and let users decide if they want only servers in their region or across all continents.

    it would also prevent issues when geo-location fails.

    I wonder if that happened when you were running a benchmark. I would appreciate a link to that report. Closest servers are found using speedtest.net API and I haven't seen it fail so far.

    Also it would be nice to include all the human readable stats at the top of the page, rather then strewn throughout.

    I assume you're talking about "I/O" tab and I can see how that may be an issue. What if I hide raw output by default and have a link there that would unfold these blocks?

  • vfusevfuse Member, Host Rep

    Good to see someone picking up after serverbear, used SB a lot when it was active!

    Here's a report for an OVH ssd-vps2 server https://serverscope.io/trials/LqP#system

  • noamannoaman Member

    @Anton said:
    Hey guys,

    I'm working on an automated benchmark tool for Linux servers and have just finished the MVP, so I figured it would be a great time to get some feedback.

    Basically, I want to eventually build a modern better version of Serverbear since it's been pretty much abandoned for the last two years.

    At this point I only have two pages:

    In the next couple months I plan to finish up comparison pages that would allow users to compare aggregated benchmark results between different VPS offers, run lots of benchmarks to collect more data (as I don't expect this thing to become popular any time soon), and add a few other benchmarks into the benchmarking script as it currently only includes dd, fio, unixbench and a bandwidth benchmark via speedtest.net.

    The benchmark tool comes as a Python script and here is what it does:

    1. Checks for dependencies and installs them if needed. Depending on the platform dependencies are either make, automake, gcc, gcc-c++, kernel-devel, libaio-devel (for yum) or build-essential,libaio-dev (for apt)
    2. Gets the location of the server from API using curl
    3. Collects some server specs using Linux tools like df and output of /proc/cpuinfo / proc/meminfo
    4. Runs benchmarking tools (fio, dd, unixbench and a custom bandwidth test)
    5. Uploads the results and returns a link to the report page

    I would appreciate all kinds of feedback, especially things that you would like to see implemented such as benchmarks or website features.

    Your. Responive. Layout needs. Fixing. .. but I like the idea and Domian

  • LiteServerLiteServer Member, Patron Provider

    I've ran a test as well. I'm happy with the way the results are being showed - good work!
    Still a few bugs though. For example, the e-mail "Your benchmark is complete" shows me a completely different IP address - one that definitely doesn't below to us :-):

    "ServerScope.io benchmark has been successfully completed for your server hosted by LiteServer (174.4.57.61) located in , Netherlands"

    Here is a test of our "OVZ-SSD-1024" plan on a loaded VPS node:
    https://serverscope.io/trials/WKK#system

  • cassacassa Member

    noaman said: Your. Responive. Layout needs.

    Because? I've never used my phone to watch a benchmark or make one

  • AntonAnton Member

    @noaman said:
    Your. Responive. Layout needs. Fixing. .. but I like the idea and Domian

    Yup. This is planned for the next iteration.

    Even though almost nobody would benchmark their server via phone, it should be an option to browse the report page using any device.

  • AntonAnton Member

    @LiteServer said:
    I've ran a test as well. I'm happy with the way the results are being showed - good work!
    Still a few bugs though. For example, the e-mail "Your benchmark is complete" shows me a completely different IP address - one that definitely doesn't below to us :-):

    Oh right! I've setup Cloudflare CDN but forgot the change the way I'm detecting IP addresses.

    Similar thing with the location string. I've just updated site to work properly when city is not detected but forgot to update the e-mail template.

    Will update that soon. Thanks!

    Thanked by 1LiteServer
  • flexflex Member

    Your service looks pretty nice.
    Please add more providers in the selection.

    soyoustart and even some german providers would be great

Sign In or Register to comment.