mhrivnak

Run catasb offline

For demo purposes, I find it handy to run openshift with catasb offline, without an internet connection at all.

Gotchas

The cluster will happily start without an internet connection. You can even provision some services successfully. But due to extensive use of nip.io, several things will fail. apb relist will fail. Using an openshift Route

mhrivnak

#!/usr/bin/env bash

# This is just where I keep my apb tool installed in a virtualenv
PYTHON=~/pythons/apb/bin/python
PATH=$PATH:~/pythons/apb/bin/

for chart in $(helm search stable | grep -v NAME | awk '{ print $1 }' | sed -e 's/stable\///'); do
    echo $chart
    mkdir -p $chart
    pushd $chart

mhrivnak

If you run this:

package main

import (
    "fmt"
    "github.com/pborman/uuid"
)

mhrivnak

import time
import unittest

import requests
from requests.auth import HTTPBasicAuth


# usage: python3 -m unittest vrtest.py

mhrivnak

#!/usr/bin/env python                                                                                    | 27                 saxutils.XMLGenerator.characters(self, 'foo')
                                                                                                         | 28                 self._delivered_checksum = True
import gzip                                                                                              | 29                 return
from xml import sax                                                                                      | 30         saxutils.XMLGenerator.characters(self, content)
from xml.sax import handler                                                                              | 31 
                                                                                                         | 32     def endElement(self, name):
                                                                                                         | 33         if name == 'checksum':
c

mhrivnak

from collections import namedtuple
from threading import Lock
import time


class Session(object):
    """ 
    This should actually be a Session object from the requests library. I'm
    just lazy and made a simple stand-in.
    """

mhrivnak

#!/usr/bin/env python
import subprocess
from urlparse import urlparse

repos = [
    # Centos
    {'name': 'centos-6-x86_64-base', 'feed': 'http://mirror.centos.org/centos/6/os/x86_64/'},
    {'name': 'centos-6-x86_64-updates', 'feed': 'http://mirror.centos.org/centos/6/updates/x86_64/'},
    {'name': 'centos-6-x86_64-extras', 'feed': 'http://mirror.centos.org/centos/6/extras/x86_64/'},

mhrivnak

def get_unit_all_or_nothing(unit):
    """ 
    Tries to download all files associated with a unit. As soon as one fails,
    it aborts and abandons what has been downloaded so far.

    I believe this is the current aggreed-upon plan.
    """
    if unit.downloaded is True:
        # nothing to do then!
        return

mhrivnak

class TestRepoUpdate(unittest.TestCase):
    @classmethod
    def setUpClass(cls):
        cls.data = StuntData()
        cls.repo_id = cls.data.create_repo(repotype='yum')
        # call repo update API here

    @classmethod
    def tearDownClass(cls):
        cls.data.cleanup()

mhrivnak

    response = None
    for x in range(self.retries + 1):
        try:
            response = session.get(
                request.url, headers=headers, timeout=(self.config.connect_timeout,
                                                       self.config.read_timeout)
            )
        except httplib.BadStatusLine:
            continue
    if response is None: