15:00:33 #startmeeting Server Working Group Weekly Meeting (2014-10-21) 15:00:33 Meeting started Tue Oct 21 15:00:33 2014 UTC. The chair is sgallagh. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:33 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:33 #chair sgallagh mizmo nirik davidstrauss stefw adamw simo tuanta mitr 15:00:33 Current chairs: adamw davidstrauss mitr mizmo nirik sgallagh simo stefw tuanta 15:00:33 #topic roll call 15:00:39 morning. 15:00:49 .hellomynameis sgallagh 15:00:50 sgallagh: sgallagh 'Stephen Gallagher' 15:00:54 .hellomynameis tuanta 15:00:55 tuanta: tuanta 'Truong Anh Tuan' 15:00:59 Hello 15:01:16 nirik: Someday, can we shorten that to .hello? 15:01:37 * junland is here 15:01:38 we could. 15:01:57 Welcome, junland. Nice to see a new face. 15:02:00 it should be a new easyfix :) 15:02:06 * danofsatx is here for observation 15:02:13 Thank you glad to be onboard with you all. 15:02:25 * sgallagh pictures danofsatx holding a clipboard and taking notes on our behavior. 15:02:39 notes? what're those? 15:02:48 .hello kevin 15:02:49 nirik: kevin 'Kevin Fenzi' 15:02:50 * danofsatx is surprised he still has a 4.0 at uni 15:03:08 Heh, that didn't take much 15:03:18 :) 15:03:29 just a short alias away. :) 15:03:36 ahoy 15:03:42 Good morning 15:04:22 .hello 15:04:23 simo: (hello ) -- Alias for "hellomynameis $1". 15:04:28 fail :) 15:04:35 oh, wait - this is server. I'm here for (limited) participation - it's the cloud meetings I attend as an interested party only ;) 15:05:03 simo: irc nick is != fas username always. ;) mine isn't for example, so easy way to get that 15:05:11 .hello dmossor 15:05:12 danofsatx: dmossor 'Dan Mossor' 15:06:07 ok, let's move to the agenda 15:06:12 #topic Agenda 15:06:22 I have two (related) items first: 15:06:33 #info Agenda Item: Fedora 21 Install Media 15:06:33 #info Agenda Item: Fedora 21 Beta Status 15:06:51 why did my mind see that as "Beta Sucks"? 15:07:04 Haha, same here. 15:07:07 ... 15:07:28 nirik: well "try" and error out :) 15:07:58 Anyone have other topics (or specific sub-topics) to add to the agenda? 15:08:12 Nope. 15:08:37 #topic Fedora 21 Install Media 15:09:01 FESCo is currently deciding on exactly how the network install media is going to function. 15:09:33 sgallagh: I was puzzled there was an issue there 15:09:36 what's the issue ? 15:09:36 We've asserted previously that network install for Server is a primary deliverable. Does anyone wish to challenge this? 15:09:43 no 15:09:47 nope. 15:09:47 Isn't it just done using PXE? 15:09:49 simo: I'll get there. 15:09:51 I think it still is aprimary deliverable 15:10:01 * simo does all his installs using network image 15:10:04 but I am perfectly fine with it being a universal one... doesn't need to only show server 15:10:13 ack 15:10:21 I also do desktop installs in VM with it 15:10:31 and it would *suck* to have to dload the DVD instead 15:10:39 Right, the issue as it stands now is that we cannot realistically have Product-only netinstall media. 15:10:51 why should it be product only ? 15:10:55 So FESCo is trying to figure out how best to handle the Server and Workstation ones. 15:11:06 I would think you select which product at install time with the netinstall 15:11:15 * danofsatx agrees with simo 15:11:29 simo: Well, at the very least, it seems counterintuitive if I put the Server netinst ISO on a cd, boot it and get prompted to install Fedora Cloud 15:11:39 (Which is what was happening up until yesterday) 15:11:48 You could select Server, but it wasn't the default 15:12:06 yeah, that part should be fixed now tho (hopefully) 15:12:39 sgallagh: it wouldn't ask ? 15:12:42 nirik: I confirmed it with last night's boot.iso 15:12:55 simo: It doesn't ask, the hub just picks the highest-ranked choice 15:13:06 oh so you want netinst images that have a different default based on which product page you downloaded them from ? 15:13:08 Which happened to be Cloud, because it and Server were both ranked as 1 and Cloud sorted first 15:13:17 simo: Well, that's one approach. 15:13:38 you need 2 images that are identical except for one line basically ... 15:13:38 The workaround we currently have is that all netinst.iso's are picking Server by default 15:13:46 (As of last night) 15:13:49 that works for me 15:13:50 perhaps we should just ask interested folks to chime in on fesco ticket? or do we really want to go over the entire thing here again? 15:14:08 nirik: from my part now I understand the problem 15:14:26 and seem clear to me (though I am biased obviously) that Server is the preimary target for netinst 15:14:33 so defaulting to server makes sense 15:15:00 it's fine for me too. 15:15:01 The only real question I was getting to was whether anyone felt strongly that Server should *not* be the default selection. 15:15:02 I agree with that too. 15:15:09 (In favor of Workstation) 15:15:11 no 15:15:28 * nirik is fine with it too. 15:15:31 sounds to me Workstation primary install media is the live d 15:15:33 *cd 15:15:45 while netinst is more for server oriented people 15:15:48 workstation use case is admin installing a bunch of machines, but then they should know how to use ks or adjust the default 15:15:54 at least that's how I always thought about it 15:16:07 Yeah, I think the WS group agrees, but I didn't want to be speaking for all of us without discussing it 15:16:11 well the default is for people "trying out" stuff 15:16:33 ifd you are doing mass installs you already have to tweak any number of things (packages and what not) so that is not the target 15:16:58 OK, so for the record: 15:17:29 Proposal: Server WG recommends that all netinstalls be universal and select Server as the default installation environment in interactive Anaconda. 15:17:43 sure. +1 15:18:19 +1 15:18:27 +1 15:18:39 +1 15:18:41 +1 15:18:45 +1 15:19:11 +1 15:19:11 (it occurs to me we could just ditch the 'workstation netinst' entirely and tell workstation people who want a netinst to use the server image, but let's talk about that elsewhere) 15:19:23 +1; “finds it quite acceptable that“ would be sufficient I guess, we don't beed to second-guess Workstation that much ☺ 15:19:37 adamw: Well, not exactly; theirs will still get their branding, just a "bug" on the default env. 15:20:18 mitr: "Server WG finds it acceptable that all netinstalls be universal and select Server as the default installation environment in interactive Anaconda." ? 15:21:02 sgallagh: Great. (The old version is fine enough with me too, hence the +1) 15:21:28 ack either way 15:21:43 #agreed Server WG finds it acceptable that all netinstalls be universal and select Server as the default installation environment in interactive Anaconda. (+8, 0, -0) 15:21:59 #info Agenda Item: Fedora 21 Beta Status 15:22:21 Whoops 15:22:22 #undo 15:22:22 Removing item from minutes: INFO by sgallagh at 15:21:59 : Agenda Item: Fedora 21 Beta Status 15:22:26 #topic Fedora 21 Beta Status 15:23:01 OK, so let's figure out if we have everything we need in place for Beta release. 15:23:14 unfortunately i still haven't gotten around to writing the test cases for the beta criteria 15:23:18 too many things on fire 15:23:22 has anyone tested them freestyle? 15:23:44 * nirik hasn't. ;( 15:23:54 I have been busy lately :( 15:24:00 Not yet. 15:24:01 I've only tested Server uptill the selecting the roles on a VM. 15:24:10 Actually, let's make that the next agenda topic. 15:24:17 danofsatx: you tested the full DC role install right ? 15:24:25 I'd like to arrange some testing of the TCs to supplement QA 15:24:52 adamw: Are you aware of any Server-specific blockers from yesterday's review? 15:25:01 no 15:25:03 I can do some TC's if you give me some document telling me what todo. 15:25:08 yes, took a few tries, and named kept crashing, but it's running 15:25:25 however, the shiny thing I absolutely haven't been working on when I should've been writing the test cases: 15:25:26 https://www.happyassassin.net/testcase_stats/21/Server.html 15:25:30 danofsatx: did you send backtraces for named ? 15:25:30 named is crashing (I think) because the service was never enabled, it was started manually by freeipa.service 15:25:36 junland: http://dl.fedoraproject.org/pub/alt/stage/21_Beta_TC4/Server/ 15:25:37 tells us that all the server validation tests were last run against Alpha RC1 15:25:52 so no-one's run even the tests we have against any Beta compose, yet. that probably needs to happen. 15:25:58 danofsatx: and why would it crash then ? that's the normal way named is started in an IPA server 15:26:26 * junland opens link 15:26:50 #info danofsatx has been running tests against the Domain Controller Role. Is encountering an issue with named. 15:27:00 I take that back, I've only used the Alpha on a VM 15:27:12 junland: once you have it, go to https://fedoraproject.org/wiki/Test_Results:Fedora_21_Beta_TC4_Server , run the tests and fill out some results :) 15:27:14 For the record, I've done at least thirty DC deploys with named working, but they've all been in a KVM guest. 15:27:22 Thank you. 15:27:23 note that with any luck, an RC1 will arrive today. 15:27:37 danofsatx: where you on baremetal ? 15:27:39 sgallagh: testing on a VM is usually fine, and a lot better than a poke in the eye 15:27:40 #action junland to jump right in with TC testing 15:27:57 adamw: I meant specifically KVM, whereas I believe danofsatx is using a different hypervisor. 15:28:00 ah. 15:28:04 Just one difference to consider 15:28:32 the other thing i've been working on which isn't writing test cases is a TUI for reporting test results, which is something so horribly wrong it's hard to explain, but might save people from figuring out how to edit wiki tables and use the {{result}} template. 15:28:43 #action danofsatx to file a bug against FreeIPA for the named start failure 15:28:50 ANYHOO, yeah, we really need to run those tests against Beta TC4/RC1 15:29:01 danofsatx: (I hope I'm not being too presumptuous there?) 15:29:23 #info we really need to run those tests against Beta TC4/RC1 15:29:35 #link https://fedoraproject.org/wiki/Test_Results:Fedora_21_Beta_TC4_Server 15:29:52 and then check all the stuff in https://fedoraproject.org/wiki/Fedora_21_Beta_Release_Criteria#Server_Product_requirements against it too. they're all *relatively* simple and folks shouldn't have too much problem figuring out how to do it without test cases written down. if anything breaks in there, file a blocker bug 15:30:19 #link https://fedoraproject.org/wiki/Fedora_21_Beta_Release_Criteria#Server_Product_requirements 15:30:24 for anyone who doesn't know, you can nominate blocker bugs at https://qa.fedoraproject.org/blockerbugs/propose_bug , or just mark them as blocking the bug 'BetaBlocker' and explain why in a comment. 15:32:01 #info for anyone who doesn't know, you can nominate blocker bugs at https://qa.fedoraproject.org/blockerbugs/propose_bug , or just mark them as blocking the bug 'BetaBlocker' and explain why in a comment. 15:33:24 OK, so to recap: 15:33:31 We need testing. Now! 15:33:49 There are currently no known blockers specifically under this group's jurisdiction 15:34:23 yep, that's basically it. i don't know of any 'systemic' issues outstanding aside from netinst; it's all just about testing the bits. Go/No-Go Meeting is Thursday, which means we hopefully don't have any blockers but if there are any we need to know *today* to have any chance of avoiding slippage. 15:34:34 Well, and the fedup situation 15:34:45 which fedup situation? :) 15:34:55 All the fedup situations :-/ 15:35:05 eh, those are just blocker bugs to me. 15:35:25 #info Go/No-Go Meeting is Thursday, which means we hopefully don't have any blockers but if there are any we need to know *today* to have any chance of avoiding slippage. 15:36:03 #info It would be appreciated if anyone with spare cycles spends some time testing Beta TC4 today. 15:37:04 * junland will do a TC this weekend + homework. 15:37:34 by this weekend we'll already have delayed the release or finalized it =) but if we delay, the testing will be handy. 15:37:38 yes 15:38:34 Ah so then before Thursday? 15:39:13 Still trying to learn the lingo here so forgive me you guys. 15:39:19 junland: No problem 15:39:38 junland: ideally, yeah. 15:39:48 Yes, the Go/No-Go decision has to be made on Thursday, hopefully based on the release candidate build happening tonight (or if needed, the one tomorrow night) 15:39:54 Alright, Ill be able to fill a TC in then. 15:39:58 first rule of fedora, there's never enough time for anything =) 15:40:11 Thank you! 15:40:16 Haha. Noted. 15:40:31 Second rule of Fedora: There's 15:40:57 #topic Open Floor 15:41:11 I have one topic for Open Floor (came up during the meeting) 15:41:56 We still don't have branding of the GUI installer images working 15:42:12 adamw: Is branding codified in the release criteria anywhere? 15:42:31 yes, but it has not been updated to say anything about products 15:42:47 what we have is about correctly identifying the release number and milestone and using the correct artwork for the release 15:43:02 i thought it did say Server in the top-right, though? 15:44:50 Well, we still have the fuzzy placeholder logo on the left 15:45:18 That's in theory supposed to be replaced by the "correct artwork", with that implicitly meaning the Product's logo 15:45:24 oh, that. 15:45:34 yeah, not in the criteria (yet) 15:46:34 #info Product GUI install media still doesn't have the Product Logo 15:46:45 adamw: Is that in the Beta or Final criteria? 15:47:15 the existing branding stuff is mostly final iirc, with a small 'version shouldn't be completety wrong' bit at alpha 15:47:48 ok, good. 15:47:57 Then this isn't a potential risk to Beta release. 15:48:12 no, i wouldn't say so. 15:48:28 #info No risk to Beta release due to branding/logo 15:48:38 Other topics for Open Floor? 15:49:22 should we have a Server WG test day ? 15:49:59 Good question 15:50:06 #topic Server WG Test Day 15:50:20 Cockpit had a test day already. 15:50:43 Do we want to schedule something specifically around Server in general or one of our technologies (like rolekit or OpenLMI)? 15:51:11 would someone have the cycles to do a proper job of it? 15:52:06 Good question. I'm not sure I can coordinate this myself. Volunteers? :) 15:52:43 unfortunately I do not think I would have the time to organize it :( 15:53:46 i don't have time to look at the existing test days, let alone run more :/ 15:53:53 I'd like to see rolekit get sufficient testing, but I think that we're probably going to need to be doing this today and tomorrow anyway or else block the release 15:54:06 So I don't think a test day is necessarily going to help there 15:54:13 I do but I don't know if it would be "proper". 15:54:22 proper? 15:54:50 junland: If you mean "I have time to organize a test day", then congratulations. It's all yours :) 15:55:17 Haha. So what do I have todo. 15:55:36 Just give me directions and Ill do my best. 15:55:44 junland: http://fedoraproject.org/wiki/QA/Test_Days/Create 15:56:03 Also http://fedoraproject.org/wiki/QA/Test_Days 15:56:22 Basically, you'd need to put together a test plan and schedule a day. 15:56:47 If you need help, ask in #fedora-server and someone will try to assist, I'm sure. 15:56:49 * junland having problems with university internet due to something in Dallas... 15:57:03 Booked marked. 15:57:25 Two questions. 15:57:26 #action junland to look into scheduling a Fedora Server Test Day 15:57:32 sorry folks, $dayjob intervened 15:57:55 I will attempt to confirm the named bug, and file BZ as needed. 15:58:02 danofsatx: Would you be willing to assist junland with test cases, since you've already done some testing? 15:58:05 What is my time frame todo this.... 15:58:17 I can try.... 15:58:27 junland: Please schedule something for before Final Freeze 15:58:37 Will do. 15:58:40 Which is 2014-11-18 15:58:51 I will try to get one of my classmates to help me out. 15:58:53 danofsatx: Thanks 15:58:59 The more, the merrier 15:59:04 * danofsatx scrolls back 15:59:24 OK, anything further on this topic? 15:59:42 We're at the top of the hour, so I'll close out the meeting in two minutes unless there's something further. 15:59:52 * adamw has nothing 16:00:09 * junland has nothing as well. 16:00:54 * tuanta has nothing. just started another meeting 16:02:15 Thanks for coming, folks 16:02:18 #endmeeting