Awesome

Overview

Comdb2 is a clustered RDBMS built on Optimistic Concurrency Control techniques. It provides multiple isolation levels, including Snapshot and Serializable Isolation. Read/Write transactions run on any node, with the client library transparently negotiating connections to lowest cost (latency) node which is available. The client library provides transparent reconnect.

Work on Comdb2 was started at Bloomberg LP in 2004 and it has been under heavy development since. More information about the architecture of the project can be found in our VLDB 2016 paper and for more information on usage please look in the Docs.

Documentation

Comdb2 documentation is included in the docs directory. It can be hosted locally with jekyll by running jekyll serve from the docs directory.

Contributing

Please refer to our contribution guide for instructions. We welcome code and idea contributions.

Quick Start

On every machine in the cluster:

1. Make sure all machines in the cluster can talk to each other via ssh.
   Copy keys around if needed.

2. Install prerequisites:

   Debian/Ubuntu

   sudo apt-get install -y  \
       bison                \
       build-essential      \
       cmake                \
       flex                 \
       libevent-dev         \
       liblz4-dev           \
       libprotobuf-c-dev    \
       libreadline-dev      \
       libsqlite3-dev       \
       libssl-dev           \
       libunwind-dev        \
       ncurses-dev          \
       protobuf-c-compiler  \
       tcl                  \
       uuid-dev             \
       zlib1g-dev
   

   CentOS 7/8

   On CentOS 8, enable the PowerTools repository first:

   dnf config-manager --set-enabled powertools
   

   yum install -y       \
       byacc            \
       cmake            \
       epel-release     \
       flex             \
       gawk             \
       gcc              \
       gcc-c++          \
       libevent-devel   \
       libunwind        \
       libunwind-devel  \
       libuuid          \
       libuuid-devel    \
       lz4              \
       lz4-devel        \
       make             \
       openssl          \
       openssl-devel    \
       openssl-libs     \
       protobuf-c       \
       protobuf-c-devel \
       readline-devel   \
       rpm-build        \
       sqlite           \
       sqlite-devel     \
       tcl              \
       which            \
       zlib             \
       zlib-devel
   

   macOS High Sierra (experimental)

   Install Xcode and Homebrew. Then install required libraries:

   brew install cmake lz4 openssl protobuf-c readline libevent
   

   To run tests, install following:

   brew install coreutils bash gawk jq
   export PATH="/usr/local/opt/coreutils/libexec/gnubin:$PATH"
   

   It is recommended to increase the open file limits, at least in the sessions which start pmux and server.

3. Build and Install Comdb2:

   mkdir build && cd build && cmake .. && make && sudo make install
   

4. Add /opt/bb/bin to your PATH

   export PATH=$PATH:/opt/bb/bin
   

5. Start pmux:

   pmux -n
   

6. (optional) Comdb2 nodes identify each other by their hostnames. If the hostname of each node isn't resolvable from other nodes, we should tell Comdb2 the full domain name to use for the current node. Most setups won't have this issue.

   Tell comdb2 our FQDN.

   vi /opt/bb/etc/cdb2/config/comdb2.d/hostname.lrl
   add current machine's name, e.g.
   hostname machine-1.comdb2.example.com
   

7. On one machine (say machine-1), create a database - this example creates a database called testdb stored in ~/db.

   comdb2 --create --dir ~/db testdb
   

   Note: the --dir PATH parameter is optional, and if it is omitted comdb2 uses a default root of /opt/bb/var/cdb2/ for creating a database directory to contain the database files, which is named as per the database name parameter; hence in this case /opt/bb/var/cdb2/testdb.
   The default root will have to be created explicitly with the desired permissions before invoking comdb2 --create for a database.
   In this quick start, we use the home directory to avoid obfuscating the key steps of the process.

8. Configure the nodes in the cluster:

   vi ~/db/testdb.lrl
   add
   cluster nodes machine-1.comdb2.example.com machine-2.comdb2.example.com
   

9. On other nodes, copy the database over:

   copycomdb2 mptest-1.comdb2.example.com:${HOME}/db/testdb.lrl
   

10. On all nodes, start the database.

    comdb2 --lrl ~/db/testdb.lrl testdb
    

    All nodes will say 'I AM READY.' when ready.

    Note: the log dir comdb2 uses by default is /opt/bb/var/log/cdb2/ If this directory does not have permissions allowing the user to create file, there will be diagnostics output such as:

    [ERROR] error opening '/opt/bb/var/log/cdb2/testdb.longreqs' for logging: 13 Permission denied

    This condition will not impact operation of the database for the purposes of this quick start.
     

11. On any node, start using the database. You don't have any tables yet. You can add them with cdb2sql Example -

    cdb2sql testdb local 'CREATE TABLE t1 {
         schema {
             int a
         }
    }'
    

    Database can be queried/updated with cdb2sql:

    cdb2sql testdb local 'insert into t1(a) values(1)'
    (rows inserted=1)
    cdb2sql testdb local 'select * from t1'
    (a=1)
    

Comdb2 Directory Contents