GithubHelp home page GithubHelp logo

jepsen-1's Introduction

Scylla

This repository implements Jepsen tests for Scylla.

You'll need a Jepsen environment, including a control node with a JVM and Leiningen, and a collection of Debian 10 nodes to install the database on. There are instructions for setting up a Jepsen environent here.

Once your environment is ready, you should be able to run something like:

lein run test -w list-append --concurrency 10n -r 100 --max-writes-per-key 100 --nemesis partition

This runs a single test using the list-append workload, which uses SERIAL reads, and LWT UPDATEs to append unique integers to CQL lists, then searches for anomalies which would indicate the resulting history is incompatible with strict serializability. The test uses ten times the number of DB nodes, runs approximately 100 requests per second, and writes up to 100 values per key. During the test, we introduce network partitions. For version 4.2, this might yield something like:

  :anomaly-types (:G-nonadjacent-realtime
                  :G-single-realtime
                  :cycle-search-timeout
                  :incompatible-order),
  ...
  :not #{:read-atomic :read-committed},
  :also-not
  #{:ROLA :causal-cerone :consistent-view :cursor-stability
    :forward-consistent-view :monotonic-atomic-view
    :monotonic-snapshot-read :monotonic-view
    :parallel-snapshot-isolation :prefix :repeatable-read :serializable
    :snapshot-isolation :strict-serializable
    :strong-session-serializable :strong-session-snapshot-isolation
    :strong-snapshot-isolation :update-serializable}},
 :valid? false}

Analysis invalid! (ノಥ益ಥ)ノ ┻━┻

Running Tests

Use lein run test -w ... to run a single test. Use lein run test-all ... to run a broad collection of workloads with a variety of nemeses. You can filter test-all to just a specific workload or nemesis by using -w or --nemesis.

As with all Jepsen tests, you'll find detailed results, graphs, and node logs in store/latest. lein run serve will start a web server on port 8080 for browsing the store directory.

Most tests are tunable with command line options. See lein run test --help for a full list of options.

Running Tests in Docker

A Docker container preconfigured to run Jepsen tests is available at tjake/jepsen on Docker Hub. Since it runs Docker inside Docker, it must be run with the privileged flag. A command like docker run -it --privileged -v /home/jkni/git:/jkni-git tjake/jepsen will start the container and attach to it as an interactive shell. Since you'll likely be running a newer version of Jepsen/C* tests than those available in the image, you'll want to share the directory containing your local Jepsen/C* clone with the container as in the example above.

Troubleshooting

If you have trouble getting Scylla to start, check the node logs, and look at service scylla-server status on a DB node.

You may need to up aio-max-nr, either on individual DB nodes, or, for containers, on the container host.

echo 16777216 >/proc/sys/fs/aio-max-nr

jepsen-1's People

Contributors

aphyr avatar enaydanov avatar kostja avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.