RLTest

Redis Labs Test Framework, allow to run tests on redis and modules on verity of environments.

Supported Environment: oss, oss-cluster, enterprise, enterprise-cluster

The framework allow you to write a test without environment specification and then run the test on all supported environment.

Install

$ pip install git+https://github.com/RedisLabsModules/RLTest.git@master

Usage:

usage: RLTest [-h] [--module MODULE] [--module-args MODULE_ARGS]
              [--env {oss,oss-cluster,enterprise,enterprise-cluster}]
              [--oss-redis-path OSS_REDIS_PATH]
              [--enterprise-redis-path ENTERPRISE_REDIS_PATH]
              [--stop-on-failure] [--verbose] [--debug]
              [--tests-dir TESTS_DIR] [--test-name TEST_NAME]
              [--tests-file TESTS_FILE] [--env-only] [--clear-logs]
              [--log-dir LOG_DIR] [--use-slaves] [--shards-count SHARDS_COUNT]
              [--download-enterprise-binaries]
              [--proxy-binary-path PROXY_BINARY_PATH]
              [--enterprise-lib-path ENTERPRISE_LIB_PATH] [--env-reuse]
              [--use-aof] [--debug-print] 
              [--use-valgrind] [--vg-verbose] [--vg-no-leakcheck]
              [--vg-options VALGRIND_OPTIONS]
              [--vg-suppressions VALGRIND_SUPPRESSIONS_FILE]
              [--interactive-debugger]
              [--debugger-args DEBUGGER_ARGS] [--no-output-catch]

module

Load a spacific module to the environment, notice that on enterprise the file should be a zip file packed with RAMP. might override by the test.

module-args

Pass arguments to the loaded module. might override by the test.

env

The environment on which to run the tests (oss,oss-cluster,enterprise,enterprise-cluster), might override by the test.

oss-redis-path

Path to the oss redis binary (default - redis-server)

enterprise-redis-path

Path to the enterprise redis binarty (default - ~/.RLTest/opt/redislabs/bin/redis-server).

stop-on-failure

Stop the tests run on failure, allows you to check what went wrong.

verbose

Increase verbosity, it is possible to write this options twice (-vv) to increase verbosity even more.

debug

Stop before each test execution and allow you to attach to any process with debuger.

tests-dir

Directory to search for tests (default - current directory).

test-name

Name of spacific test function to run.

tests-file

File inside the test_dir to search for tests (if not specified the framework searches in all files)

env-only

Not running any tests, just setup the environment and keep it up for manual tests.

log-dir

Directory on which logs will be written to (default - ./logs).

use-slaves

Setup the environment with slaves. might override by the test.

shards-count

On cluster, setting the amount on required shards. might override by the test.

download-enterprise-binaries

Downloading the enterprise binaries before running the tests and save it under ~/.RLTest/.

proxy-binary-path

Enterprise proxy binary path (default - ~/.RLTest/opt/redislabs/bin/dmcproxy)

enterprise-lib-path

Enterprise libraries requires for run (default - ~/.RLTest/opt/redislabs/lib/)

env-reuse

Reuse the existing env between tests. Notice that if some test requires a different env setting then the current env will be taken down (and a new one will be setup) regardless this parameter value.

use-aof

Using aof instead of rdb, might override by the test.

use-valgrind

Run redis under valgrind (assuming valgrind is installed on the machine).

valgrind-suppressions-file

Path to valgrind suppressions (not mandatory).

interactive-debugger

runs the redis on a debugger (gdb/lldb) interactivly. debugger interactive mode is only possible on a single process and so unsupported on cluste or with slaves. it is also not possible to use valgrind on interactive debugger. interactive debugger direcly applies: --no-output-catch and --stop-on-failure. it is also implies that only one test will be run (if --inv-only was not specify), an error will be raise otherwise.

debugger-args

arguments to pass to the debugger.

no-output-catch

all output will be written to the stdout, no log files.

Configuration File

By default the framework search for configuration file on the current directory. The configuration file name is: config.txt. It is possible to specify different configuration file on command line using the '@' prefix, for example:

RLTest @myConfig.txt # search for myConfig.txt configuration file

The configuration file format is the same as the command line argument, i.e : '--< param_name > < param_val >'.

It is also possible to comment a spacific lines in the configuration file using '#'.

Example:

-vv
--clear-logs
#--debug

Test Example

from RLTest import Env
import time


class testExample():
    '''
    run all tests on a single env without taking
    env down between tests
    '''
    def __init__(self):
        self.env = Env()

    def testExample(self):
        con = self.env.getConnection()
        con.set('x', 1)
        self.env.assertEqual(con.get('x'), '1')

    def testExample1(self):
        con = self.env.getConnection()
        con.set('x', 1)
        self.env.assertEqual(con.get('x'), '1')
        self.env.assertFalse(True)  # check failure

    def testExample2(self):
        con = self.env.getConnection()
        con.set('x', 1)
        self.env.assertEqual(con.get('x'), '1')


# run each test on different env
def test_example():
    env = Env()
    con = env.getConnection()
    con.set('x', 1)
    env.assertEqual(con.get('x'), '1')


def test_example_2():
    env = Env()
    env.assertOk(env.cmd('set', 'x', '1'))
    env.expect('get', 'x').equal('1')

    env.expect('lpush', 'list', '1', '2', '3').equal(3)
    env.expect('lrange', 'list', '0', '-1').debugPrint().contains('1')
    env.debugPrint('this is some debug printing')


def test_example_3():
    env = Env(useSlaves=True, env='oss')
    con = env.getConnection()
    con.set('x', 1)
    con2 = env.getSlaveConnection()
    time.sleep(0.1)
    env.assertEqual(con2.get('x'), '1')