NetworkManager/contrib/rh-bkr

These are utility scripts to interact with the beaker installation
of Red Hat to run tests of NetworkManager.


./bkr.py:
=========

Quick-Setup:

yum install kobo python-nitrate python-termcolor
yum install beaker-client

$ cat ~/.nitrate
[nitrate]
url = https://tcms.engineering.redhat.com/xmlrpc/

$ cat ~/.beaker_client/config
HUB_URL = "https://beaker.engineering.redhat.com"
AUTH_METHOD = "krbv"

Find $DISTRO_NAME at
  https://beaker.engineering.redhat.com/distrotrees/?simplesearch=RHEL-7.1



Examples:

  kinit
  export TEST_URL="http://download.eng.brq.redhat.com/scratch/$USER/NetworkManager-rhel-7.tar.gz"
  ./bkr submit --no-test -j job.xml -r 'file://NetworkManager*.rpm'
  ./bkr submit --no-test -j job.xml -r 'jenkins://429/.*rpm'
  ./bkr submit --no-test -j job.xml -r NetworkManager.rpm
  ./bkr submit -r NetworkManager.rpm -r http://somewhere.com/NetworkManager-glib.rpm

Requirements:

  - install and configure 'beaker-client'
    http://beaker-project.org/docs/user-guide/bkr-client.html
    Also adjust your /etc/beaker/client.conf

  - configure your kerberos authentication properly and ensure,
    that your user is authenticated to schedule jobs on beaker.

Important: run kinit otherwise the script won't work.



Currently the script only supports one command: 'submit' to submit
a job to beaker. See the available options with:

    ./bkr.py submit --help

To submit a job, you must provide a beaker XML job configuration.
This file must be passed to the submit command with the '-j/--job'
argument.

The job file can contain placeholders such as $NAME that will be replaced
by the script before submitting the job. Currently the following placeholders
are supported:

  - $RPM_LIST a whitespace separated list of all RPM URLs

  - $$ a single $ symbol

If a placeholder cannot be found, it will try to look into the environment
variables to find a match. So, you can set additional variables by setting
the environtment. As last attempt, it will lookup for a hard coded list of
default values. If the name for a placeholder cannot be substituted, it
will be replaced by the empty word "". You can wrap the name in braces to
separate it from the following text (${NAME}).


You can specify any number of RPMs to the script using the '-r/--rpm'
argument. These names are expected to be an URL that is reachable by the
beaker instance, so that it can download the RPM from there.

Currently the following types are supported:

  - http:// and https:// parameters are passed on unmodified

  - file://[glob] a file glob for a local file. The files will be uploaded to
    the public_html directory of file.brq.redhat.com using rsync/ssh. Afterwards
    the url to http://file.brq.redhat.com/~username/filename will be used. For this
    you must have a kerberos ticket. The username is parsed from the ticket
    name as returned by klist.

  - jenkins://[BUILDNR](/[URIREGEX]): the jenkins build server builds
    RPM packages for RHEL-7.0 and stores them as artifacts. By using this
    URI, it will resolve the URLs for the build number BUILDNR.
    If URIREGEX is specified, this regex is used to restrict the number of
    found files. Otherwise, a default regex is used that relevant packages.

  - brew://[ID](/[URIREGEX]): fetches URLs from
    https://brewweb.engineering.redhat.com/brew/buildinfo?buildID=ID, i.e. from
    the officially build packages.

  - brewtask://[ID](/[URIREGEX]): fetches URLs from
    https://brewweb.engineering.redhat.com/brew/kinfo?taskID=ID, i.e. the builds
    when doing a scratch build with rhpkg.

  - repo:<baseurl>[/<regex-wildcard>.rpm]: fetches from a RPMMD (Yum) repository
    Useful with official Fedora repositories or COPRs

You can provide arbitrarily many -r flags and they will be joined together to
form $RPM_LIST.


Unless called with --no-test, no files are really uploaded and the beaker job is not
scheduled. Instead it prints only, what would be done and creates the processed job xml
file.