16:01:00 <gundalow> #startmeeting Networking meeting
16:01:00 <zodbot> Meeting started Wed Aug 24 16:01:00 2016 UTC.  The chair is gundalow. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:00 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:01:00 <zodbot> The meeting name has been set to 'networking_meeting'
16:01:08 <gundalow> #chair Qalthos privateip
16:01:08 <zodbot> Current chairs: Qalthos gundalow privateip
16:01:27 <gundalow> Hey eveyone, who's around?
16:01:34 * caphrim007 raises hand
16:01:35 <gundalow> caphrim007: hi
16:01:40 * mhite waves
16:01:45 <privateip> good day all
16:01:46 * Qalthos 🌊🌊
16:02:18 <gundalow> Agenda as always https://github.com/ansible/community/issues/110
16:02:48 <gundalow> #topic Tracking Network OS versions
16:04:06 <caphrim007> gundalow: i've added some topics
16:04:11 <gundalow> For now on we will be including details of the version of the network appliance that's been tested in when doing bug fixing
16:04:14 <gundalow> caphrim007: Thanks :)
16:05:30 <gundalow> privateip: You got the example, I can't see it now
16:05:56 <privateip> https://github.com/ansible/ansible-modules-core/pull/4523
16:05:56 <gundalow> For example a string like: Tested with VyOS 1.7
16:06:01 <gundalow> Thanks
16:06:06 <privateip> np
16:06:11 <gundalow> We also ask for that information when people raise bugs
16:07:11 <gundalow> Just as freeform text, the aim is we can 1) Track what a PR has been tested against, which helps if bugs are found later on 2) Lets us know what people are using
16:07:15 <gundalow> privateip: anything you'd like to add?
16:08:30 <stianvi> gundalow - this should also include hardware version/type too?
16:08:44 <gundalow> stianvi: excellent question. I think that would be useful
16:08:53 <stianvi> i.e. qfx5100 junos 15.1R4
16:09:36 <gundalow> One would hope that the APIs are common across appliances, however that may not be the case
16:10:13 <privateip> hw is nice to have but the os will be the req element
16:10:24 <privateip> in some cases hw just doesn't make sense
16:11:54 <gundalow> It's a new thing, we will try it out and see how it works. Easy to adjust in the future
16:12:42 <gundalow> Any other questions/thoughts on that?
16:12:47 <mhite> Seems reasonable.
16:13:49 <gundalow> #topic 2.2 code freeze
16:14:11 <gundalow> #info Ansible 2.2 is coming. https://groups.google.com/forum/#!topic/ansible-devel/Ce5uNA5TnJ0
16:14:27 <gundalow> #n preparation for major releases, we will be instituting a "feature freeze" before the first release candidate. For 2.2, the last day to add features will be Monday, September 5th.
16:14:31 <gundalow> During this period, no new features will be added to the devel branch as we stabilize the code and prepare to create the stable-2.2 branch. This freeze will last at least 2 weeks, but could last up to a month or more depending on any bugs we find related to the new features we've created in the devel branch.
16:14:41 <gundalow> ..
16:15:27 <gundalow> So the Network group, like the rest of the Core Engineering Team are busy finishing off their features for that
16:15:29 <caphrim007> does that date apply to -extras too?
16:16:04 <gundalow> caphrim007: -extra will be ~ 2 weeks after that
16:16:08 <caphrim007> k
16:16:19 <gundalow> I see you have more modules for us :)
16:16:46 <stianvi> so just for the core and core-modules?
16:16:46 <caphrim007> yeah i wanted to get them in before 2.2 window closed
16:16:54 * privateip labels caphrim007 "the module machine"
16:17:00 <gundalow> +!
16:17:03 <gundalow> +1
16:17:10 <linuxdynasty> +1
16:17:42 <mhite> @caphrim007 is definitely beast mode on the F5 modules right now.
16:17:56 <gundalow> stianvi: I think we will allow some modules-core changes in, though maybe not as long a window as modules-extras
16:18:33 <stianvi> thanks gundalow
16:19:22 <gundalow> #topic Ansible 2. Network Modules updates
16:19:27 * gundalow hands over to privateip
16:19:34 <privateip> thx
16:19:46 <privateip> just a quick update for those here (and those reading at home later)
16:20:00 <privateip> you might have noticed a flurry of commits this week
16:20:12 <privateip> we are starting to get the network 2.2 refactor merged back into devel
16:20:27 <privateip> as of now the following are code complete (still in testing)
16:20:35 <privateip> ios_config, ios_commnad, ios_facts
16:20:41 <privateip> nxos_config, nxos_command
16:20:50 <privateip> eos_command, eos_config
16:21:00 <privateip> iosxr_command, ios_config (coming today)
16:21:08 <privateip> iosxr_command, iosxr_config** (coming today)
16:21:16 <privateip> vyos_command, vyos_config
16:21:21 <privateip> still to come
16:21:32 <privateip> ops_command, ops_config (openswitch)
16:21:37 <privateip> junos_command, junos_config
16:21:44 <privateip> facts modules for all
16:22:08 <privateip> i would suggest starting serious testing next week
16:22:21 <privateip> please file issues for anything you find
16:22:33 <privateip> lots of goodness added to both
16:22:48 <caphrim007> do those command modules use...a webservice api?
16:23:00 <privateip> where available
16:23:00 <caphrim007> or is it a cli via paramiko?
16:23:03 <caphrim007> or something else?
16:23:21 <privateip> all support cli, we have implemented api calls for those platforms that support them
16:23:53 <privateip> the two biggest changes: *_command will now enforce no config commands and *_config will now load a config from disk
16:24:07 <caphrim007> ahh ok
16:24:07 <privateip> this will ultimately allow us to deprecate *_template
16:24:41 * privateip takes a bow and returns to gundalow
16:24:48 * gundalow wakes up
16:24:59 <gundalow> Any other questions on that?
16:25:08 <privateip> haha i have a habit of doing that to people :)
16:25:30 <gundalow> nah, was just realising other stuff we needed to do and making notes :(
16:25:54 <gundalow> #topic caphrim007's F5 modules
16:26:32 <caphrim007> i sent in several. they've been baking in our other repo for a while now
16:27:29 <gundalow> privateip: Are you happy for me to merge any that have "shipit" LABEL
16:27:38 <privateip> yep
16:27:51 <caphrim007> i think i've tried to address any comments on them from mhite, gundalow, etc
16:28:06 <gundalow> Cool, I'll review after meeting and do so
16:28:10 <gundalow> caphrim007: in https://github.com/ansible/ansibullbot/blob/master/MAINTAINERS-EXTRAS.txt
16:28:12 <caphrim007> i added another pr for a bigip_facts enhancement a couple moments ago
16:28:30 <gundalow> Should the whole of network/f5 just be allocated to you?
16:28:43 <caphrim007> yeah probably, at a minimum
16:28:52 <caphrim007> mhite: want to be on that list too?
16:29:04 <caphrim007> (sees he's there)
16:29:11 <gundalow> being on that list means you will get a ping when someone raises a PR
16:29:22 <caphrim007> that's fine
16:29:29 <gundalow> `network/f5/: mhite caphrim007`
16:29:55 <gundalow> caphrim007: did you take over bigip_gtm_wide_ip.py from perzizzle?
16:30:16 <caphrim007> i think they mentioned they were ok with that?
16:30:21 <caphrim007> so I'm ok with that
16:30:28 <gundalow> cool, I'll update that
16:30:46 <caphrim007> i'll need to make sure that it's part of the f5-sdk before i move forward on it
16:31:00 <caphrim007> so it may not be ready for 2.2
16:34:14 <gundalow> caphrim007: https://github.com/ansible/ansible-modules-extras/pull/2777/files did you decide to remove the singular option ntp_server, and just use ntp_servers (with s)
16:35:04 <caphrim007> ....y......e....s.....
16:35:14 * caphrim007 goes to look at his tests
16:35:25 <gundalow> YAY TESTS
16:35:37 <gundalow> I think the docs just need trimming don
16:35:39 <gundalow> down*
16:36:09 <caphrim007> yeah it looks like i decided against the two different params
16:36:16 <caphrim007> lemme update that
16:36:24 <gundalow> aye, one option is much better
16:38:12 <gundalow> #topic Open Floow
16:38:14 <gundalow> #topic Open Floor
16:38:24 <gundalow> What else do people have?
16:38:37 <gundalow> Qalthos: Anything?
16:39:37 * Qalthos looks for the tab
16:40:12 <gundalow> #topic Provide options for SSH key checking https://github.com/ansible/ansible/pull/15736
16:40:16 <gundalow> #link https://github.com/ansible/ansible/pull/15736
16:40:23 * gundalow hands over to Qalthos
16:40:52 <Qalthos> Basically, as it is now, using shell to provide SSH connections doesn't actually do any sort of host checking
16:41:28 <Qalthos> It pretends to, but doesn't actually consult any known_hosts files, making it somewhat pointless
16:42:53 <Qalthos> This change adds three options for using shell... 'ignore', which preserves the current behavior, 'strict', which checks user and system host files and rejects hosts it doesn't know, and 'loose', which does the thing most people want
16:43:45 <Qalthos> ie, allow hosts, but check host files and write back to ~/.ssh/known_hosts (unless I misunderstand paramiko docs)
16:44:40 <privateip> +1 for loose as the default for 2.2
16:45:12 <gundalow> caphrim007: mhite stianvi What are your thoughts?
16:45:26 <gundalow> + anyone else lurking in here :)
16:45:46 <mhite> I'm ok with loose
16:46:03 <skg_net> +1 loose as default as well
16:46:08 <caphrim007> i concur
16:46:09 <Qalthos> sweet
16:46:14 <gundalow> skg_net: hi :)
16:46:45 <skg_net> @gundalow Hi
16:46:50 <Qalthos> unless anyone has any objections, I will change the default to loose and merge today, then
16:48:29 <gundalow> cool, thanks Qalthos
16:48:37 <gundalow> #topic Open Floor
16:48:41 <gundalow> Anyone got anything else?
16:49:09 <skg_net> is there a way for Ansible to persist the ssh connection
16:49:17 <skg_net> I tried ssh_args = -o ControlMaster=auto -o ControlPersist=60s but no luck
16:50:05 <privateip> not today when using connection=local (ie cli over ssh)
16:50:31 <skg_net> Thanks
16:50:36 <privateip> for devices that offer a true linux environment then the connection can be changed to ssh in which case the control channel is persisted
16:51:27 <skg_net> ok got it..let me play around and get back
16:53:35 * gundalow makes a note to add that to the docs (when I update them)
16:55:31 <gundalow> Anything else, otherwise I'll close the meeting
16:56:13 <skg_net> Monday I will be raising PRs for DELL modules
16:56:34 <gundalow> skg_net: ah yes
16:57:29 <gundalow> skg_net: ping me here, or "@gundalow" in the PR and I'll look at it
16:57:34 <gundalow> #closemeeting
16:57:43 <gundalow> #meetingclose
16:57:48 <skg_net> ok will do it
16:57:51 <gundalow> #endmeeting