Schaufel - shovel data around

Copyright (c) 2017, Robert Abraham

https:/adjust/schaufel

schaufel aims to be a swiss army knife for moving data. It can take data from list-like sources and insert them into list-like sinks. In other words, it can take data from a redis list and insert it into kafka or vice versa.

Producers and consumers of all types can be run in parallel.

schaufel tries to be efficient but primitive. Its queueing can exceed 150000 messages/second but is upper bound by lock contention (as it works on single elements instead of pipelining).

It compiles on linux, freebsd and netbsd, but work is in progress to remove all gnuisms. It does compile with libmusl.

schaufel is licensed under the MIT license.

Features

• kafka support (through librdkafka)
• redis support (through hiredis)
• file sink/source support
• postgres sink support
• json dereferencing

Documentation

• Introduction and manual in INTRODUCTION.md.
• man pages man 1 schaufel and man 5 schaufel.conf

Installation

Build from source

Requirements

gcc/clang
GNU make
pthreads
librdkafka
hiredis
libpq (postgres)
libjson-c
a libc that supports hcreate_r/tdestroy

Building

Makefile based build

make
make install

Autotools based build

autoreconf -fi
./configure
make

Tests

* make test
    run testsuite (do not run in parallel if using old Makefile)
* scripts/debug.sh
    run testsuite with Werror and debug symbols
* scripts/sanitize.sh
    run sanitizer and valgrind testsuite