Observer class hangs when passed an interface that isn't up about pathspider

[irl]

@britram I have no ideas for this. Do you have any ideas?

[britram]

I think the answer here is "don't do that, then". Probably involves rooting around in python-libtrace. Let's come back around and dig into this one later.

On 01 Jul 2016, at 15:39, Iain Learmonth [email protected] wrote:

@britram https://github.com/britram I have no ideas for this. Do you have any ideas?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub #16 (comment), or mute the thread https://github.com/notifications/unsubscribe/ABy_1KQFofI1CkjLPqOq3z5gLeSZt2lAks5qRRiJgaJpZM4IEK1V.

[irl]

Changes the milestone on this one then, we can deal with this later.

[britram]

We'll need to have a look at python-libtrace anyway for accelerating TFO. I'm talking to Nevil next week about getting his code into github, and what would be necessary to add features to it.

[irl]

In order to fully support multihoming I'll be using pyroute2 (interface to netlink) which means we can add a test to check to see the interface is up before starting.

[irl]

netlink code is now merged, and this is no longer affecting pathspider, though is still a problem in python-libtrace.