16:09:30 <privateip> #startmeeting ansible networking 16:09:30 <zodbot> Meeting started Wed Jan 25 16:09:30 2017 UTC. The chair is privateip. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:09:30 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:09:30 <zodbot> The meeting name has been set to 'ansible_networking' 16:09:38 <privateip> sorry for the delay all 16:09:40 <privateip> good morning 16:09:56 <privateip> gundalow is tied up so i will moderate today 16:10:48 <resmo> hi 16:11:18 <st8less> fyi, I'm @djustice on NTC 16:11:25 * bcoca goes to basement to untie gundalow 16:11:45 <privateip> #topic open floor 16:11:57 <privateip> open floor today 16:12:26 <privateip> please not everyone at once though 16:12:29 <privateip> ;) 16:13:17 <st8less> My first meeting, but I'm game. PR 20262 16:13:25 <privateip> cool 16:13:31 <privateip> do you have the link handy? 16:13:38 <st8less> https://github.com/ansible/ansible/pull/20262 16:13:44 <privateip> thx 16:14:54 <gundalow> hello 16:15:03 <gundalow> sorry, got tied up in another meeting 16:15:03 <privateip> not sure why the need for the setters 16:15:20 <privateip> hi gundalow 16:15:27 <privateip> we heard bcoca untied you :) 16:15:55 <st8less> nxos_vxlan_vtep_vni inherits a CommandLine object and tries to modify private properties. 16:16:40 <privateip> i dont see that in the code ... do you have th elinum handy? 16:16:51 <privateip> wow ... how about the linenum handy? 16:17:02 <st8less> I may have quoted the wrong class, just a sec. Looking 16:17:25 <privateip> sure 16:17:27 <st8less> Sorry, ConfigLine class. line 222 16:18:13 <privateip> in netcfg? 16:18:28 <st8less> No nxos_vlan_vtep_vni 16:18:35 <privateip> oops 16:20:13 <privateip> is this the place? 16:20:27 <privateip> https://github.com/ansible/ansible/blob/devel/lib/ansible/modules/network/nxos/nxos_vxlan_vtep_vni.py#L222 16:22:21 <st8less> In that file, the instance is created on line 217. But on line 231, it tries to access self.indent and there is not getter for that property in netcfg 16:22:48 <st8less> Another instance is created on 248 16:23:34 <st8less> On lines 235 and 250, the module tries to set parents which has no setter method in netcfg. 16:23:56 <privateip> oh i see whats going on here 16:24:08 <st8less> There is absolutely no way that module worked out of the box, no offense. 16:24:19 <privateip> netcfg has changed since 2.2 was released 16:24:24 <privateip> your right this is broke 16:24:38 <privateip> i dont think adding the setters is the right change though 16:24:50 <privateip> i think we need to update nxos_vxlan_vtep_vni 16:25:10 <privateip> is ggabriele around? 16:25:37 <st8less> There were other bugs I fixed in the module as well, the netcfg stuff is just local to the 'add' method. I'm open to suggestions on fixing it. 16:25:38 <gundalow> privateip: nop, he was on networktocode earlier, he said he could'nt make IRC today 16:26:03 <gundalow> though if you @mention him in issues he will repond 16:26:43 <privateip> ok i think the right path forward is to get the nxos_ modules off this common block of code from 2.1 16:27:22 <privateip> lets hold the PR until ggabriele can weigh in 16:27:39 <privateip> agreed? 16:27:48 <gundalow> +1 16:28:50 <privateip> okay updated PR with comments 16:29:42 <privateip> other PRs there are questions about? 16:29:59 * gundalow added a few bits on the agenda 16:30:33 <gundalow> most were PRs, there was a question about .updates 16:30:49 <privateip> lets tackle the updates one 16:30:54 <privateip> #topic return key updates 16:31:32 <privateip> there is a discrepancy between return keys when we merged template into config 16:32:12 <privateip> for cli based modules i'm recommending that all cli based modules return a commands key wich is the list of commands sent to the device if a change is necessary 16:32:13 <gundalow> https://github.com/ansible/ansible/issues/19382 16:32:28 <privateip> this effectively means updates = commands moving forward for consistency 16:32:32 <privateip> for all network modules 16:32:43 <privateip> especially as we get into declarative modules 16:32:57 <bcoca> privateip: make prposal, not just network, we have diff modules with 'action', 'details', 'logs', etc we need unified way for this 16:33:09 <privateip> point taken 16:33:19 <bcoca> ^ would be nice to have 'one way' 16:33:23 <privateip> agreed 16:33:25 <privateip> will take that action 16:34:18 <privateip> #action privateip to write up proposal on standardizing return keys 16:34:53 <bcoca> http://docs.ansible.com/ansible/common_return_values.html <= here is where we keep the documented stdrds on return 16:35:34 * privateip should RTFM more often 16:36:46 <bcoca> ^ it used to be in my head until recently 16:36:58 <bcoca> wrote it down for ... scale 16:37:16 * privateip doesn't want to read bcoca's head, its dark and cold and scary 16:37:26 <bcoca> youknowit! 16:38:07 <privateip> light day today ... other topics for discussion? 16:38:12 <privateip> #topic open floor 16:38:34 <ganeshnalawade> I had one issue to discuss 16:38:43 <ganeshnalawade> https://github.com/ansible/ansible/issues/19107 16:38:44 <privateip> hi ganeshnalawde 16:39:24 <ganeshnalawade> Hi privateip 16:40:58 <privateip> hmm i thought this was fixed 16:42:05 <privateip> if im reading this right the issue is a delete statement is trying to execute when there is no corresponding section to delete 16:42:15 <privateip> is that how you are reading it ganeshnalawade? 16:42:20 <ganeshnalawade> yes 16:42:49 <ganeshnalawade> if statment in not found on junos device it throws a warning and continues execution 16:43:06 <ganeshnalawade> i think with Ansible also need to have same beahviour 16:43:32 <ganeshnalawade> because of this issue user need add dummy configuration in playbook 16:43:52 <privateip> got it 16:44:13 <ganeshnalawade> need to check the error severity in <rpc-reply>..if it it 'warning' need to continue Ansible run 16:44:43 <privateip> also do you know if once the error is hit does the system stop processing any further statements? 16:45:21 <ganeshnalawade> if error severity is 'error' it stops execution 16:45:47 <privateip> ok i need to test this as well to understand the behavior 16:45:56 <ganeshnalawade> ok 16:46:18 <ganeshnalawade> Netconf protocol defines different levels of error severity 16:46:27 <privateip> yep 16:46:53 <privateip> currently writing a netconf connection plugin that will solve lots of these nagging issues 16:47:10 <ganeshnalawade> ok 16:47:11 <privateip> in fact just committed the first bits this morning 16:47:53 <privateip> just committed in the issue 16:48:30 <ganeshnalawade> got it...I think we need to parse rpc-reply and add a check of <error-severity> tag 16:48:46 <ganeshnalawade> that's all i had 16:48:51 <privateip> thanks 16:48:59 <privateip> last call for topics for today 16:50:00 <privateip> going twice 16:52:04 <privateip> thanks all see you next week 16:52:07 <privateip> #endmeeting