15:02:52 #startmeeting FESCO (2020-05-04) 15:02:52 Meeting started Mon May 4 15:02:52 2020 UTC. 15:02:52 This meeting is logged and archived in a public location. 15:02:52 The chair is decathorpe. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:02:52 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:02:52 The meeting name has been set to 'fesco_(2020-05-04)' 15:02:57 #meetingname fesco 15:02:57 The meeting name has been set to 'fesco' 15:03:02 #chair nirik, ignatenkobrain, decathorpe, zbyszek, bookwar, sgallagh, contyk, mhroncok, dcantrell 15:03:02 Current chairs: bookwar contyk dcantrell decathorpe ignatenkobrain mhroncok nirik sgallagh zbyszek 15:03:09 #topic Init Process 15:03:10 .hello2 15:03:11 .hello2 15:03:11 bcotton: bcotton 'Ben Cotton' 15:03:13 zbyszek: zbyszek 'Zbigniew Jędrzejewski-Szmek' 15:03:16 morning 15:03:47 .hello2 15:03:48 bookwar: bookwar 'Aleksandra Fedorova' 15:03:55 hi 15:04:25 good evening 15:04:41 .hello2 15:04:42 sgallagh: sgallagh 'Stephen Gallagher' 15:06:26 ignatenkobrain won't make it today. should we wait for contyk? dcantrel / dcantrell doesn't seem to be online. 15:07:08 contyk is on PTO 15:07:22 decathorpe: dcantrell was just giving a tech talk; he might still show up 15:07:32 FWIW, I didn't have time to study up on the one ticket we have on the agenda 15:07:44 But I wouldn't necessarily wait for him 15:07:48 okay, then let's start 15:07:59 #topic #2372 F33 Self-contained Change: Network Time Security 15:08:04 .fesco 2372 15:08:07 decathorpe: Issue #2372: F33 Self-contained Change: Network Time Security - fesco - Pagure.io - https://pagure.io/fesco/issue/2372 15:08:44 oh, hello dcantrell :) 15:08:48 #chair dcantrell 15:08:48 Current chairs: bookwar contyk dcantrell decathorpe ignatenkobrain mhroncok nirik sgallagh zbyszek 15:08:50 .hello2 15:08:52 dcantrell: dcantrell 'David Cantrell' 15:08:56 hi 15:09:43 doesn't look like much has happened here since last week 15:10:02 So, I have a proposal: 15:11:00 Proposal: "FESCo approves Network Time Security support for Fedora 33, but does not approve it to be enabled by default. Please write up instructions for enabling it that we can include in release notes." 15:11:50 Hmm, but is an approval necessary to enable an non-default feature? 15:12:15 If I read it correctly, it still involves replacing some of the innards. 15:12:38 i'd argue it's not generally required, but sgallagh's proposal makes FESCo's position clear about what is permitted for this release 15:13:16 (and the technical details of this specific proposal may or may not fall under the "still needs a change proposal" category) 15:13:25 "introduce new smth" can still be filed as a change, and change should be approved by fesco 15:13:32 I like the proposal ... and I think this is how other changes have been introduced in the past? make optional in one release, possibly enable in the next one if it works? 15:13:41 * nirik would rather here more details... 15:14:48 OK, if we treat the change as "Update to 4.0 and enable NTS while at it", it makes sense to have this is as a change, and it makes sense for us to approve it. 15:15:57 well, the questions are all still around 'enable NTS' no? 15:16:03 like, pointing to what? 15:16:19 Also, what about the anaconda parts? 15:16:36 chrony is already updated and supports nts in rawhide. 15:17:34 so this is already implemented, minus the "enabled by default" part? lovely 15:19:02 well, the chrony part. No idea on installer... and it's not enabled in the default config. 15:19:30 proposal: invite change owners to next weeks meeting to answer questions and/or ping them to answer questions in ticket/list? 15:19:42 +1 to nirik 15:20:13 +1 to nirik. some clarifications would be good ... 15:20:21 +1 to nirik 15:21:38 bookwar, sgallagh, dcantrell : vote? 15:22:07 0 15:22:20 i think clarification at the meeting would be hard, i'd rather have discussion in the ticket 15:22:36 it seems change owner is ready to talk, but we need our questions 15:23:31 so +1, but let's make sure we start the conversation then, rather then just postpone it for one more week 15:23:36 +1 to nirik 15:23:58 bookwar: good point 15:24:01 I guess re-reading it they want us to answer that... 15:24:09 ie, should it be default or not. 15:24:16 and pointing to what 15:25:01 I guess we answered that already, don't enable it by default ... but I don't think users would be thrilled to rely on Cloudflare for this 15:25:52 * nirik is now reconsidering sgallagh's proposal. :) 15:26:02 so, relying on cloudflare on default system seems to be a "no go" for me 15:26:27 * nirik agrees. 15:26:37 ++ if that's the only option, it must be opt-in 15:26:45 so, we really don't have anything to point it to... so this is most 'if you want to enable it it's available now' 15:26:55 if we can not make it default in some other way - then we can just approve it as sgallagh said 15:27:46 It does say "best option seems to be Cloudflare" which implies other options exist. 15:28:19 But in general, I think I'd rather see us spend a cycle figuring this out. So I'm still in favor of my earlier proposal 15:28:24 I don't personally want to get into the business of running nts servers. That seems... not alined with the fedora infrastructure goals/mission. 15:28:37 i believe that "no change" changes are important too, they need a place in release notes, and discussion on mailing lists and tracking, so i am for +1 on he change with no defaults 15:29:04 so you want to vote on sgallagh's original proposal? 15:29:09 yeah, +1 to sgallagh's proposal... seems the best we can do. 15:29:23 * decathorpe resetting my vote count 15:29:33 It may come to this, but then I think the Change page needs to be adjusted to clearly describe what is happening and what is not happening. So either way, we're back to discussion with the Change owners. 15:30:00 I agree with zbyszek 15:30:04 So I'm -1 to approving the change right now, because the descripition is confusing/unclear. 15:30:09 I don't like us to vote on something that is not the change 15:30:18 ok, so lets ask them to adjust ? 15:30:19 not if the change proposal is not updated 15:30:25 * sgallagh nods 15:30:27 works for me 15:30:49 proposal: ask change owners to update the change to the "no changing defaults" variant and vote in the ticket as soon as the update is done 15:31:14 * decathorpe resets vote count again 15:31:18 +1 15:31:45 +1 15:32:02 +1 15:32:57 +1 15:33:02 +1 to bookwar 15:33:33 zbyszek: vote? 15:33:47 -0, I think this could use some more discussion. 15:35:00 E.g. PEERNTP=, is this a thing? I cannot find the documentation anywhere. 15:35:22 #agree Ask change owners to update the Change proposal to not change the default configuration (+6, 1, -0) 15:35:31 zbyszek: will you ask that in the ticket? 15:35:46 bookwar: yep 15:36:09 hm. did I mess up the zodbot syntax? 15:36:24 decathorpe: should be good, there is no response 15:36:39 good. moving on 15:36:44 it doesn't respond to agree, but it works fine. 15:36:51 #topic #2381 F33 System-Wide Change: systemd-resolved 15:36:55 .fesco 2381 15:36:59 decathorpe: Issue #2381: F33 System-Wide Change: systemd-resolved - fesco - Pagure.io - https://pagure.io/fesco/issue/2381 15:37:20 BTW, https://bugzilla.redhat.com/show_bug.cgi?id=809367 is for PEERNTP. 15:37:33 sgallagh: you tagged this with meeting, I guess because of the -1 vote? 15:37:52 Right, anything with a -1 goes to the meeting. 15:38:14 I count (+2, 1, -1) from votes in the ticket. 15:38:25 do we want to discuss this during the meeting? 15:38:40 I reached out today to some of the folks on the Red Hat Security Team to get their opinions on systemd-resolved because I have been out of that space long enough to be unsure of the state of my understanding. 15:39:07 The responses that I got back were largely not in favor, so I'm effectively proxying that opinion. 15:39:34 huh... thats somewhat vuage. :( 15:39:34 sgallagh: is that somehting that should be dicussed on the devel list rather than a fesco meeting? 15:39:47 I'm with sgallagh on this one. The idea to me sounds fine for Fedora, but the implementation seems clunky 15:39:50 with details and time to read the information asynchronously 15:39:52 * nirik switched to it a while back and the only thing I have hit so far is https://bugzilla.redhat.com/show_bug.cgi?id=1823480 15:40:54 The major concern is the modifications to nsswitch.conf which are not trivially reversible. 15:41:11 (e.g. by simply doing `systemctl disable systemd-resolved.service` 15:41:38 Proposal: post feedback on the devel list and restart discussion 15:41:45 +1 15:41:47 decathorpe: +1 15:42:53 +1 15:43:11 sgallagh: actually doing 'systemctl disable systemd-resolved.service' will effectively disable the changes to nsswitch.conf. 15:43:17 sgallagh: it uses a user one I thought? 15:43:19 I'm leaving my -1 on the ticket for now in the interests of not having it get auto-approved. 15:43:28 nirik: What do you mean? 15:43:31 +1 to more discussion in any case. 15:43:50 +1 to more disucssion, that is always good 15:43:51 "(a) modifying authselect's user-nsswitch.conf template, if authselect is in use, or (b) directly modifying /etc/nsswitch.conf otherwise. " 15:44:29 bookwar: vote? 15:44:32 we can take this elsewhere 15:44:37 +1 15:44:48 oh, thats the template. sure. 15:45:16 #agree Post feedback on the devel list and restart discussion (+7, 0, -0) 15:45:33 #topic Next week's chair 15:45:46 Igor has volunteered to chair next week's meeting since he couldn't make it today, any objections? 15:45:47 ignatenkobrain 15:45:55 fine by me 15:46:34 (I don't think we need to ack or vote on next week's chair) 15:46:38 #action ignatenkobrain will chair next meeting 15:46:40 (at least we never did) 15:46:51 #topic Open Floor 15:46:53 . 15:47:48 should we discuss how to handle the new fesco elections question changes not to select the questions ourselves in a fesco ticket? 15:48:00 I can give a short datacenter move update. 15:48:23 (I also have something about the security policy update.) 15:49:34 nirik: short update sounds good 15:49:44 mhroncok can give longer updates after that :) 15:49:45 mhroncok: where do you propose to brainstorm the questions? 15:49:52 decathorpe: unless it's "everything's on fire" ;-) 15:50:08 bookwar: on the devel mailing list, in the thread where we ask people to chip in 15:50:14 bcotton: well, it's short and non-actionable, so we could just move on 15:50:22 RDU2 datacenter: we had a dead switch, hopefully replaced soon. Hopefully communishift will be able to come back up later this week. IAD2 (new virgina datacenter): We finally have network access. We spent the weekend configuring management interfaces for new hardware. There's some more to go and then we will start bringing up instances there. So, currently, we are on schedule still to move the week of june 8th hopefully. 15:50:29 they might have feedback to our proposals, but they are not aware about them,because they don't follow the fesco tracker 15:50:56 nirik++ 15:50:56 mhroncok: Karma for kevin changed to 9 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 15:50:58 nirik++ 15:51:05 mhroncok: good point, so let's reserve ticket for tracking the change, while keep discussion on a mailing list 15:51:22 mhroncok: do you want to reply to the devel thread? 15:51:38 what I don't knwo is how do we select the questions at the end 15:51:49 I don't think we need to have more and more quations, or do we? 15:52:06 decathorpe: I'd rataher if the proposals are shared there by their authors 15:52:08 nirik: will you announce communishift on devel mailing list as soon it is up? or where should i look 15:52:31 bookwar: yes, can send to devel-announce 15:52:33 (let's finish nirik's topic first, this is ugly) 15:52:52 mhroncok: i am done, sorry for that 15:52:59 let's focus on questions topic 15:53:21 sgallagh, decathorpe: please, could you bring your porposals to devel list? 15:53:30 Will do 15:53:33 will do 15:53:37 thanks 15:53:43 everybody: how will we select the questions and when? 15:53:56 should be complete by 19 May 15:53:58 #action decathorpe and sgallagh will respond to the devel list concerning FESCo election questionnaire 15:54:18 IMHO we need something votable by fesco at least a week in advance 15:54:21 that is 12 May 15:54:30 so brainstorm on mailing list, summarize the outcome of mailing list discussion some day and post the summary to the ticket, vote on each question separately 15:54:33 that is basically next week meeting 15:54:46 bookwar: so basically, just adding more questions? 15:55:00 ~4 questions with most of the votes get into the interview 15:55:17 I'm all for dropping one of the existing questions, two of them are pretty similar 15:55:31 bookwar: I'd for example argue that a "Who are you" question should not count 15:56:02 also "4 questions with most of the votes get into the interview" will likely generate conflicts, do we range vote? 15:56:37 mhroncok: to be honest, i would look into list first, then see where to put a line. I am not sure if we need a strict policy on this 15:56:38 mhroncok: maybe let's just collect the questions and have someone do de-duplicatio by hand. 15:56:40 my point is: if we want change, now is the time to decide what the framewrok for the change will be, becasue we have a deadline 15:57:11 Proposal: Collect questions, bcotton has the final say on which are used 15:57:20 we can certainly have a volunteer, who would collect the ideas, feedback and than propose a final list as a proposal 15:57:50 Either option works. 15:58:00 deadline should affect the day when we cut the discussion and do a summary, i think, but the we can just look at it 15:58:02 sgallagh: if bcotton is able to work on that, I think I like your idea, except that fesco should probably ack it 15:58:18 then* 15:58:23 it really depends on bcotton's availability 15:58:25 i can do that 15:58:26 * nirik is fine with giving bcotton more work. ;) 15:58:30 sgallagh: assignig volunteers works for me :) 15:58:38 bcotton++ 15:58:38 mhroncok: Karma for bcotton changed to 8 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 15:58:43 bcotton++ 15:58:43 decathorpe: Karma for bcotton changed to 9 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 15:58:43 mhroncok: I disagree on general principle. I don't think FESCo members should get to decide which questions determine if they get to continue holding their seat :) 15:58:52 but definitely agree with having FESCo approve my results 15:59:00 sgallagh: I see your point and I share the concern 15:59:16 sgallagh: but I also disgree that one person should get to decide which questions determine... 15:59:43 IMHO this should technically be council's decision 15:59:58 but I am afraid that having that involved, we won't make the deadline 15:59:59 sgallagh: i think we are going too political. do you expect we fight for it to keep our places? 16:00:13 :) 16:00:21 * sgallagh hides his pistol behind his back. "No?" 16:00:32 I mean if yes, we can certainly do a formal community range vote, but i am not sure we really need that 16:00:37 The interview may include parts which don't answer one of the specific questions, but talk about other things too. So I don't think we need to overthink this. 16:00:51 ok, let me summarize this 16:01:06 I trust bcotton, so consider me an automatic +1 on his proposed questions 16:01:09 Sufficient? 16:01:23 bcotton: no pressure 16:01:26 :-D 16:01:47 1) send proposals to devel. 2) bcotton collects the ideas and feedback, submits a new list for fesco ack. 3) fesco acks, automatically getting a +1 from sgallagh 16:01:49 i think the concerns are valid but also solving for a problem we haven't experienced in my....experience 16:02:10 mhroncok: sounds good 16:02:10 +1 for mhroncok's proposal 16:02:14 mhroncok: +1 16:02:17 Who does 1? 16:02:30 zbyszek: anybody, or nobody 16:02:32 devel list subscribers I guess :) 16:02:38 zbyszek: have ideas? do share them 16:02:49 Right, but we need an #action to have somebody start the discussion lest we forget. 16:02:52 zbyszek: no ideas? bcotton presents the current list 16:02:53 bcotton: As election coordinator, do you mind starting the thread with the existing suggestions from the ticket? 16:02:58 +1 to the proposal 16:03:09 i accept the #action 16:03:43 +1 to the proposal 16:03:51 alright, let me get this right ... 16:04:14 side note: I also trust bcotton but I just deem "one person selects the questions" as not very good idea on principle (sorry bcotton) 16:04:52 mhroncok: no offense taken :-) 16:05:05 #agree Collect more questions on devel list, ask FPgM to curate, and FESCo will approve questions (+5, 0, -0) 16:05:18 #action bcotton to ask for questions on the devel list 16:06:02 any objections to those two ^ items? 16:06:15 * bcotton has no objections 16:06:56 no objections 16:07:10 anything else? otherwise I'll close the meeting in a few minutes. 16:07:47 did we get nirik's datacenter update finished? 16:08:03 I didn't have anything more, unless there were more questions? 16:08:11 i thought there were some questions that got tabled while we were discussion elections 16:08:26 * nirik reads back 16:08:28 * mhroncok waits until this is calrified 16:09:03 I dont see any off hand, but please re-ask if anyone had any 16:09:57 * mhroncok considers that topic clarified now 16:10:06 I only see bookwar's question and that was answered, so I think we're good 16:10:10 I just wanted to say that the security policy doesn't seem to have a consensus on devel and the discussions stopped a long time ago. Not sure if we want to just keep the status quo (a policy that we don't follow) or take ay other action 16:10:41 * decathorpe makes a note to read up on that discussion 16:11:36 policy on CVE bugs you mean? 16:12:32 mhroncok: can you open a FESCo ticket for that so we don't lose track of it again? 16:13:45 decathorpe: roger that, action me 16:13:48 nirik: yes, that one 16:13:59 thanks 16:14:14 #action mhroncok to open fesco ticket about Security Policy 16:14:17 yeah, it would be good to do something there. 16:14:50 anything else for the Open Floor? 16:16:03 guess not :) thanks everybody. 16:16:06 #endmeeting