Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    Timofey Barmin
    @timofey-barmin
    500 workers each
    Urjit Singh Bhatia
    @urjitbhatia
    awesome! thanks for the answer :thumbsup:
    what is the recommended max workers per node for EC2? say c4.large instances
    running 5000 total on 2 nodes caused pool crashes
    (using the recommended AMIs)
    Timofey Barmin
    @timofey-barmin
    I would say it depends on your worker and worker script
    Urjit Singh Bhatia
    @urjitbhatia
    hmm, I am using the vernemq mqtt load testing script
    the funny thing is that CPU, Memory etc on the nodes looked healthy
    Timofey Barmin
    @timofey-barmin
    it should not crash
    Urjit Singh Bhatia
    @urjitbhatia
    is there any documentation about how many ports I can open for the default AMIs etc? I guess I need to do some more test runs and tweak the nodes themselves
    does this point anywhere? exit with reason reached_max_restart_intensity in context shutdown ?
    Renat Idrisov
    @parsifal-47
    it depends on your linux settings, could be checked with
    echo /proc/sys/net/ipv4/ip_local_port_range
    more information on this:
    for example, to increase port range:
    echo 1024 65535 > /proc/sys/net/ipv4/ip_local_port_range
    reached_max_restart_intensity it does not say much, for some reason it has crashed, OTP was trying to restart it, it got crashed again until it reached maximum intensity. You need to look above this message or at some different logs to find out what was the root cause
    Urjit Singh Bhatia
    @urjitbhatia
    Thanks @parsifal-47 that and some other tweaks helped me get 24k parallel connections for my load test
    Adrian Lewis
    @aidylewis
    Ever thought about using Elixir as a scripting language and if you did do, do you think you'd get more traction?
    Adrian Lewis
    @aidylewis
    How is this product different from the other Erlang load tool: Tsung?
    Urjit Singh Bhatia
    @urjitbhatia
    imo: better UI, cloud support etc
    Adrian Lewis
    @aidylewis
    Do you have any documentation on the Python client?
    Renat Idrisov
    @parsifal-47
    Adrian Lewis
    @aidylewis
    @parsifal-47 Yes. Thanks.
    Do you have any plans to implement HTTP/2 - mainly server push and multiplexing?
    Adrian Lewis
    @aidylewis
    V
    Do you still use Graphite or a time-series database?
    Do you support SSL client cert authorisation, which we use for APIs and pre-live testing?
    Renat Idrisov
    @parsifal-47
    @aidylewis HTTP/2 is interesting, but I have no idea about exact plans, please create an issue and we'll keep you up to date with this information. We don't use graphite, metric data is stored in own format. We support for SSL certs for server, here is the information: https://github.com/machinezone/mzbench/blob/master/doc/deployment.md#protocol
    liquidharmonic
    @liquidharmonic
    Hi Everyone. I'm looking to load test a clustered VerneMQ server and it's doc lead me to MZBench and the vmq_bench worker plug in.
    I've searched around but there doesn't seem to be a lot of community support (i.e. stack overflow posts). I was just wondering if anyone could point me in the right direction in terms of support.
    Note. I have vmq_bench working locally when I rsync a local version of vmq_bench.
    Renat Idrisov
    @parsifal-47
    @liquidharmonic Hi, I've seen some documentation here: https://github.com/erlio/vmq_mzbench
    if it is not what you looking for, feel free to post issues and ask questions
    we'd be glad to help
    liquidharmonic
    @liquidharmonic
    Hi @parsifal-47 . Thanks for the link. If I run into any issues I'll open an issue https://github.com/erlio/vmq_mzbench.
    Does the load and RAM metrics in MZbench represent the load and RAM on the MZBench API server or the nodes MZBench allocates? I'm quite green when it comes to erlang so it's difficult for me to grasp the source code. Thanks
    Urjit Singh Bhatia
    @urjitbhatia
    @liquidharmonic I have successfully tested a vernemq cluster with mzbench - lmk if you have questions
    (did it on aws)
    Renat Idrisov
    @parsifal-47
    @liquidharmonic it is the consumption on allocated nodes, feel free to ask your questions
    liquidharmonic
    @liquidharmonic
    Thank you @urjitbhatia for offering your help. I am currently setting up all the frameworks that have MQTT scale testing locally. I've successfully set up MZBench and LocustIO.
    @parsifal-47 I'll ask questions as I come across them. Thank you so much already
    liquidharmonic
    @liquidharmonic
    @parsifal-47 How efficient is mzbench at creating pools of workers? I assume it does it very well because erlang was built for optimizing concurrency.
    Renat Idrisov
    @parsifal-47
    @liquidharmonic shortly -- yes, I have no numbers now, but at some point we tested 1M workers, it was less then a minute as far as I remember, I could check :)
    liquidharmonic
    @liquidharmonic
    thanks @parsifal-47. Random question, was your git name inspired by King Arthur or Ready Player One? :smile:
    liquidharmonic
    @liquidharmonic
    @parsifal-47 @urjitbhatia I posted an issue with string interpolation machinezone/mzbench#124 with an example of using vmq_benchmark with 100,000 workers subscribing to unique topics using the iterator.
    @urjitbhatia I'm curious about the scenarios you created when you tested on the cloud.
    Urjit Singh Bhatia
    @urjitbhatia
    @liquidharmonic I was trying to load a lot of concurrent users and then scale their req/sec to find benchmark vernemq+our h/w selection
    I used the aws cloud provider to launch workers
    Renat Idrisov
    @parsifal-47
    @liquidharmonic sorry for late answer, it is more about Wagner :)
    raghav922
    @raghav922
    Hello
    can anyone help me in mzbench issue?

    ```mzbench requires erl_terms, docopt and requests

    These packages can be installed using your preferred package manager
    or using pip, typically:

    sudo pip install -r requirements.txt