data/
      some data for testing (e.g. package lists)
etc/
      validation.yaml - example configuration (AWS credentials, keys, ...)
examples/
      validation examples
hwp/
      hardware profiles
scripts/
      valid_runner.py - main validation runner
valid/
      source code

Example: valid_runner.py --data examples/example_rhel63_58_all_x86_64.yaml

Data file is a yaml-encoded list. Example:

examle_datafile.yaml: - ami: ami-bafcf3ce arch: x86_64_bug914737 product: Fedora region: eu-west-1 version: '18'

Mandatory fileds:

• ami: EC2 AMI id
• region: EC2 region name
• arch: name of hwp file (/usr/share/valid/hwp/.yaml)
• product: product name ("RHEL", "Fedora", ...)
• version: product version (e.g. "18", "6.3", ...)

Optional fields:

• name: name tag for testing
• itype: Instance type ("hourly", "access")
• disable_stages: disable specified stages
• enable_stages: enable specified stages only (overrides disable_stages)
• disable_tags: disable specified tags
• enable_tags: enable specified tags only (overrides disable_stages)
• disable_tests: disable specified tests
• enable_tests: enable specified tests only (overrides disable_stages)
• repeat: repeat testing with the same instance N times

Data file is being updated by hwp data!

HWP file is also a yaml-encoded list. Example:

x86_64_hvm_kernel.yaml: - arch: x86_64 cpu: '4' ec2name: m3.xlarge memory: '14000000' virtualization: hvm - arch: x86_64 bmap: - {delete_on_termination: true, name: /dev/sda1, size: '15'} - {ephemeral_name: ephemeral0, name: /dev/sdf} cpu: '32' ec2name: cr1.8xlarge memory: '244000000' virtualization: 'hvm'

Mandatory fields:

• arch: hardware architecture ("x86_64", "i386")
• ec2name: EC2 instance name ("t1.micro", "m3.xlarge", ...)
• cpu: expected numer of CPUs
• memory: expected amount of memory
• virtualization: virtualization type ("hwp", "paravirtual")

Optional fields:

• bmap: block device map
• userdata: userdata string

There are several useful tools available:

• valid_bugzilla_reporter.py - examine result, report to bugzilla (--test option is useful as well)
• valid_debug_run.py - run one specified test on existing instance

Python-patchwork library is required. You can get it here: https://github.com/RedHatQE/python-patchwork or you can alternatively download prebuilt RPM here: https://rhuiqerpm.s3.amazonaws.com/index.html

Validation can be done in client/server mode as well. The communications are secured with HTTPS (default port is 8080). You can create certificates using 'valid_cert_creator.py' script. You should use real hostname you'll be connecting to, it will be checked during ssl negotiations. Client certificate and key are generated in /etc/valid/ directory. On server side: 'valid_runner.py --server' (or use provided systemd valid.service)

On client side: valid_client.py --cert 'certfile' --key 'keyfile' --host 'hostname' --add 'datafile' [ --emails 'comma-separated email list'] you'll get the transaction id.

and then valid_client.py --cert 'certfile' --key 'keyfile' --host 'hostname' --get 'transaction id'

There are some examples in valid/testing_modules directory. The test is a class which looks like this:

valid/testing_modules/testcase_xx_testname.py:

from valid.valid_testcase import *


class testcase_xx_testname(ValidTestcase):
    # applicable stages
    stages = ["stage1"]
# testing tags
tags = ["default"]
# applicable setups
applicable = {"product": "(?i)RHEL|BETA", "version": "5.*|6.*"}

    def test(self, connection, params):
    # doing testing

'connection' is a patchwork.connection.Connection object. 'params' is a data line united with hardware profile and runtime information (so you can use something like params["memory"], params["product"], ...) Setting "stages" and "tags" is mandatory. Don't forget to include your new test in valid/testing_modules/init.py. Tests are executed in alphabetical order with repetition count and stage name as prefix (e.g. 01stage1testcase_xx_testname).

Prebuilt RPMs are available here: https://rhuiqerpm.s3.amazonaws.com/index.html