ip matching is too aggressive

[Kevin W Monroe]

In utils.py, we check for /etc/hosts entries and update them if found, append them if not found. The ip matching re is too aggressive, and will match if an ip is a subset of another (e.g. 10.0.0.123 will match when scanning for 10.0.0.12).

This is an easy fix -- just add a whitespace char to the re, but all apache-hadoop core charms will need an updated charmhelpers.tgz.

[Kevin W Monroe (kwmonroe)]

Fixed in charmhelpers r442:

http://bazaar.launchpad.net/~bigdata-dev/charm-helpers/framework/revision/442

[Cory Johns (johnsca)]

Now that https://pypi.python.org/pypi/jujubigdata is available, I'd like to switch to that instead of updating charmhelpers, as it will allow us to change our resources entry to this:

jujubigdata:
    pypi: jujubigdata>=1.0.0,<2.0.0

This will give us the ability to make backwards compatible bug fixes and improvements without updating all of the charms individually, and if we need to make a change that would break the existing charms, we can still do that by increasing the major version number, and then updating & fixing the charms, so that deployments of the older versions will continue to work with the older jujubigdata version.

IOW, it gives us much better control over our versioning and avoids excessive churn on the charms.

[Cory Johns (johnsca)]

Also, since we'd no longer be relying on our own contrib changes in charmhelpers, we have the option of moving back to using the upstream version directly and not having to maintain our own fork.