13:03:44 <dtroyer> #startmeeting openstackclient 13:03:44 <openstack> Meeting started Thu Jun 2 13:03:44 2016 UTC and is due to finish in 60 minutes. The chair is dtroyer. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:03:45 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:03:48 <openstack> The meeting name has been set to 'openstackclient' 13:04:02 <tangchen> It is night for me. :) hahaha 13:04:30 <rtheis> good night too :) 13:04:51 <tangchen> :) 13:05:49 <dtroyer> Shall we get started? 13:05:56 <dtroyer> #topic releases 13:06:01 <tangchen> let's go 13:06:18 <dtroyer> We did the 2.5.0 release last week, the last planned release before 3.0 13:06:46 <dtroyer> I wanted to hold off making any breaking merges until we had a bit of time to make sure 2.5.1 was not immediately necessary 13:07:08 <dtroyer> we seem to be clear of that 13:07:21 <dtroyer> anyone know of anything otherwise? 13:07:43 <rtheis> dtroyer: I may have one... 13:08:16 <rtheis> I just found an issue late yesterday that I think is related to --enable-beta-commands 13:08:47 <rtheis> Somehow, --enable options are getting ignored on commands now 13:09:15 <rtheis> I need to investigate further as to why this is. Any ideas? 13:09:22 <dtroyer> oh good grief…yes 13:09:42 <dtroyer> global option names bite again 13:10:12 <dtroyer> I forget the exact setting, but argparse will swallow up substring matches sometimes 13:10:39 <dtroyer> that's another reason we use —os on most global options 13:10:39 <rtheis> oh boy 13:11:09 <rtheis> so is it best to just change to --os prefix for this? 13:11:37 <dtroyer> possibly, or something similar 13:11:55 <dtroyer> even just —os-beta-commands? 13:12:07 <rtheis> sure 13:12:14 <rtheis> I like that 13:12:24 <rtheis> I'll open a bug and work on a fix today. 13:12:31 <dtroyer> ok, thanks 13:12:52 <rtheis> I didn't expect argparse to have such a "feature" 13:12:59 <dtroyer> I have a feeling though that we'll need to jump to 2.6.0 based on what has already merged…I will review 13:13:11 <rtheis> ok 13:13:51 <dtroyer> IIRC argparse does this when combined with… tab complete maybe? 13:14:03 <dtroyer> no, not tab complete… dang, I forget exactly what it is 13:15:30 <rtheis> that's ok 13:15:32 <dtroyer> anything else release-wise? This means we need to still hold off on the bigger breaking things for a while, probably next week 13:15:43 <rtheis> I don't have anything else 13:15:52 <tangchen> none for me 13:16:17 <dtroyer> #topic reviews 13:16:34 <dtroyer> So I was wanting to go through the stuff we have been holding back here 13:16:49 <dtroyer> we can still look at the list, but wait on them 13:17:28 <tangchen> Are we going to do the ip refactor now ? after 2.5.0 13:17:48 <dtroyer> yes, that set is one in this category 13:17:52 <tangchen> I mean this series. https://review.openstack.org/#/c/300388/ 13:17:54 <tangchen> OK 13:18:21 <dtroyer> also Navid's KSA and it's follow on reviews from Alvaro 13:18:39 <dtroyer> and Henry's role assignment 13:19:19 <rtheis> ok 13:19:44 * dtroyer is looking at the list from last week 13:20:24 <dtroyer> so the only things I am unsure about is what may be new since last Thursday 13:21:24 <dtroyer> other than release-type issues, any reviews we need to talk about? 13:21:55 <rtheis> Looking for opinions on command object for https://review.openstack.org/#/c/319522/ 13:22:13 <rtheis> This is for neutron rbac policies 13:22:47 <dtroyer> wow, yeah, yuck 13:23:13 <rtheis> Similar for https://review.openstack.org/#/c/320604/ which is start of OSC neutron plugin 13:23:32 <rtheis> that is, plugin for network advanced services 13:24:09 <rtheis> no rush on these, just wanted to bring them up for attention 13:24:20 <dtroyer> sure, thanks, I had not seen them 13:24:44 * dtroyer was sleeping much of yesterday, too much holiday weekend maybe :( 13:25:05 <rtheis> :) 13:25:27 <dtroyer> interestingly enough, the recent ML thread about changing the *aaS release model brings up the near-deadness of at least two of those projects 13:26:33 <dtroyer> back to the RBAC bit though, I don't think we can assume that 'rbac' is network-only in the long term 13:26:53 <dtroyer> but yes, it's a bad name all on its own 13:27:36 <rtheis> yeah, it seemed similar to qos in my mind 13:27:50 <dtroyer> how does 'network policy' fit? it feels like it might still be too ambiguous 13:28:18 <rtheis> network has qos policies 13:28:21 <dtroyer> qos has a better case though, as it is the common term in the networking world. like IP 13:28:24 <rtheis> and rbac policies 13:28:31 <dtroyer> ah, ok 13:28:51 <rtheis> we have "volume qos" today 13:28:58 <dtroyer> what role does the R in RBAC refer to? 13:29:53 <dtroyer> users or servers? 13:30:19 <rtheis> trying to sort that out now 13:30:20 <tangchen> I think it is users. 13:30:25 <dtroyer> I have always thought of it as user-related but don't see how that work in a neutron context 13:31:11 <rtheis> it is users actions allowed for network objects enforced on a project 13:31:46 <rtheis> I haven't played with it much 13:32:22 <dtroyer> Identity uses plain role here 13:32:23 <tangchen> me nither... 13:32:33 <dtroyer> so a) it owns that object, but b) doesn't use rbac 13:32:41 <dtroyer> name 13:33:10 <dtroyer> ok, let's keep the conversation going in the reviews 13:33:24 <dtroyer> I'll leave some comments after the meeting 13:33:30 <rtheis> thanks dtroyer 13:33:42 <tangchen> thanks. 13:34:09 <dtroyer> tangchen: in https://review.openstack.org/#/c/321977/ I noticed you added an exception for when no option (—enable|—disable) is present 13:34:16 <dtroyer> I thought we started removing those complaints 13:34:52 <dtroyer> thinking that if the user does not specify a change, that is not an error 13:35:30 <tangchen> Well, that makes sense. So, a warning msg ? 13:36:02 <tangchen> I think we should let the user know nothing has changed. 13:36:03 <dtroyer> maybe… I still lean toward nothing though 13:36:39 <dtroyer> maybe an info message, that will be silent unless -v is set 13:37:35 <tangchen> I think nothing will be a little strange. The command is not complete. but it looks like it has been executed correctly. 13:37:53 <dtroyer> but it has been executed correctly 13:38:02 <dtroyer> it did nothing correctly 13:38:03 <dtroyer> :) 13:38:35 <tangchen> Well, I'd vote for a warning, but info is also ok for me. 13:38:45 <dtroyer> set commands usually have more options, if this one grows more we would need to remove the specific warning about --enable 13:38:56 <dtroyer> because then it would not be an issue 13:40:03 <tangchen> sorry, what specific warning about --enable ? 13:41:14 <dtroyer> if we did the same thing to 'user set', for example, you would have a warning if —enable or —disable was not present 13:41:40 <tangchen> Oh, I see 13:41:41 <dtroyer> which is totally acceptable 13:41:47 <tangchen> just like service set 13:42:10 <tangchen> OK 13:42:20 <tangchen> I'm OK with nothing output. 13:42:36 <tangchen> Then we need to fix some other commands. 13:42:43 <dtroyer> yes we do 13:42:50 <tangchen> OK, I'm on it 13:42:56 <dtroyer> thanks 13:43:15 <dtroyer> any other reviews? 13:43:30 <rtheis> none from me 13:43:48 <tangchen> none for me, but I have two things need some advice. 13:43:55 <tangchen> please give me 5 min in the end 13:44:02 <dtroyer> ok, will do 13:44:06 <dtroyer> #topic bugs 13:44:37 <dtroyer> I saw https://bugs.launchpad.net/python-openstackclient/+bug/1587927 and never considered that before, we have another thing to sort out 13:44:37 <openstack> Launchpad bug 1587927 in python-openstackclient "Quotas command is not plugable" [Undecided,New] 13:45:14 <dtroyer> that is a good example of when we want a plugin to be able to modify another command 13:45:21 <dtroyer> other times we don't want that 13:45:24 <dtroyer> fun times ;) 13:45:42 <rtheis> that is interesting 13:46:16 <dtroyer> I played with something similar by adding —ram —cpu and —disk to server create a while back via a plugin 13:46:42 <dtroyer> but that technique isn't viable for actual use because it depends on install order 13:47:11 <dtroyer> also, https://bugs.launchpad.net/python-openstackclient/+bug/1580870 is asking for what looks to be nova-net commands? 13:47:11 <openstack> Launchpad bug 1580870 in python-openstackclient "Fixed IP reserve/unreserve/show in computeV2" [Undecided,New] - Assigned to aohuanxuan (huanxuan-ao) 13:47:29 <dtroyer> if they are part of Compute 13:48:01 <rtheis> nova-net is back to deprecated, isn't it? 13:48:34 <dtroyer> let's see… it is an even year, and even month, an even day, so I think yes? 13:48:49 <rtheis> :) 13:49:14 <tangchen> If they are just for nova, yes. 13:49:43 <dtroyer> If we do those for Network and adding them for Compute is near-trivial, then maybe. Otherwise I think at this point we should have a Really Good Reason to add more nova-net-only commands 13:50:03 <dtroyer> I'll ask that in the bug… 13:50:19 <rtheis> agreed 13:50:24 <tangchen> OK 13:50:49 <dtroyer> other bugs to raise? 13:50:55 <tangchen> non for me 13:51:00 <rtheis> nothing else 13:51:15 <dtroyer> #topic open discussion 13:51:25 <dtroyer> what is on your ming, tangchen? 13:51:35 <tangchen> https://blueprints.launchpad.net/python-openstackclient/+spec/log-usage 13:51:44 <tangchen> Need some advice for this 13:52:01 <tangchen> I'm often confused the way to use a logger. 13:52:12 <tangchen> I can see 3 loggers in OSC 13:52:24 <tangchen> Shall we give some rules on how to use them ? 13:53:00 <dtroyer> good idea, yes 13:53:28 <rtheis> https://github.com/openstack/python-openstackclient/blob/master/openstackclient/common/command.py#L37 13:53:34 <dtroyer> I can see places where all of those make sense though, the rule will probably be based on what type of message is being reported 13:53:41 <rtheis> I thought this was an attempt at that 13:54:10 <dtroyer> rtheis: I would agree with that. 13:54:27 <tangchen> sorry, I don't quite understand 13:54:31 <dtroyer> the default for commands seems like it should be using self.log() 13:54:40 <tangchen> what do you mean richard ? 13:55:07 <dtroyer> self.app.log() should be for things that are not command-specific 13:55:17 <dtroyer> so volume's use there probably needs to be changed 13:55:25 <tangchen> yes 13:55:50 <tangchen> OK, I can start the work from volume. 13:56:07 <tangchen> then we can discuss more in the reviews, and give a rule in doc finally. 13:56:10 <dtroyer> the module-level LOG is likely the least necessary one for reporting 13:56:19 <dtroyer> sounds good 13:56:51 <tangchen> one more thing 13:56:57 <tangchen> http://docs.openstack.org/contributor-guide/writing-style/word-choice.html 13:57:02 <rtheis> tangchen: just wondering if we can expand what is already done so commands have less to worry about, but I realize more is needed as you point out 13:57:04 <tangchen> someone showed me this... 13:57:24 <tangchen> OK, Richard. :) 13:57:42 <dtroyer> ooooh! nice! 13:58:22 <tangchen> What I want to ask is, are we going to accept patches like this ? 13:58:31 <tangchen> maybe...openstack->OpenStack 13:58:36 <tangchen> too trivial, I think 13:58:40 <dtroyer> for fixing docs? 13:58:44 <tangchen> yes 13:58:48 <rtheis> depends 13:58:52 <dtroyer> we'll accept them, but I really don't want to see one for each 13:58:53 <tangchen> there was one today. 13:58:59 <tangchen> yes 13:59:25 <tangchen> She abandoned finally. And maybe will give a big one at last. 13:59:29 <dtroyer> I used to do sweeps through fixing a bunch of help and docs all at once, broken by APi or something to keep it from getting too big 13:59:49 <tangchen> OK, I see. 13:59:59 <tangchen> no question for me today. :) 14:00:16 <dtroyer> but style and form do matter, so we do want to fix these things 14:00:52 <rtheis> agreed 14:01:01 <tangchen> OK 14:01:01 <rtheis> nothing else from me 14:01:10 <dtroyer> ok, we're out of time 14:01:13 <dtroyer> thanks guys! 14:01:22 <dtroyer> #endmeeting