01:39:47 #startmeeting Workstation WG (2020-07-14) 01:39:47 Meeting started Wed Jul 15 01:39:47 2020 UTC. 01:39:47 This meeting is logged and archived in a public location. 01:39:47 The chair is cmurf. Information about MeetBot at http://wiki.debian.org/MeetBot. 01:39:47 Useful Commands: #action #agreed #halp #info #idea #link #topic. 01:39:47 The meeting name has been set to 'workstation_wg_(2020-07-14)' 01:39:49 #meetingname workstation 01:39:49 The meeting name has been set to 'workstation' 01:39:51 #chair cmurf 01:39:51 Current chairs: cmurf 01:40:09 #topic Rollcall 01:40:11 #info present wg members: cmurf, Neal, aday, mclasen, mcatanzaro, petersen, otaylor, langdon, tpopela 01:40:13 #info regrets wg members: kalev 01:40:15 #info present guests: Davide Cavalca, Alex, salimma, feborges, Josef 01:40:17 #topic Approve minutes for 7 Jun meeting 01:40:19 #link https://meetbot.fedoraproject.org/fedora-meeting-2/2020-07-08/workstation.2020-07-08-05.06.log.html 01:40:21 #agreed no objections 01:40:23 #topic Announcements 01:40:25 #info Neal: has proposed a btrfs talk for flock 01:40:27 Michael: abrt has been broken for 3 or 4 weeks due to the data center move. Retrace server broken, but local backtraces seem to be broken too. It's expected to be fixed in 3 (?) weeks. 01:40:29 #info: Abrt retrace server is down for several weeks, affects local backtraces 01:40:31 #action Michael to create a workstation issue 01:40:33 #info Davide: start of a btrfs getting started guide 01:40:35 it's currently in a Google Doc. Will hopefully land in Fedora Quick Docs soon. 01:40:37 #topic Follow-ups & Status reports 01:40:39 #topic Nano by default 01:40:41 #link https://pagure.io/fedora-workstation/issue/103 01:40:43 #info F33 change approved by FESCo. What are the next steps? 01:40:45 #action Neal to implement the change 01:40:47 #topic Swap on zram 01:40:49 #link https://pagure.io/fedora-workstation/issue/127 01:40:51 #info F33 change approved by FESCo and we voted in favour. Changes on-track. 01:40:53 Next steps in the ticket - https://pagure.io/fedora-workstation/issue/127#comment-664475 01:40:55 Chris: most of what needs to happen has happened. 01:40:57 #topic general topics 01:40:59 #topic btrfs by default 01:41:01 #link https://pagure.io/fedora-workstation/issue/153 01:41:03 #info Davide and Neal have been working on image build infra with btrfs subvolumes 01:41:05 #info Test day happened last week 01:41:07 #link https://testdays.fedorainfracloud.org/events/88 01:41:09 One interesting bug relatd to automatic partitioning of multiple disks, this is going to get fixed by Btrfs upstream soon. Other results were as expected. More tests likely needed. 01:41:11 FESCo approved the change proposal +8 -1. 01:41:13 The WG never voted on the proposal, and there's the outstanding proposal from Owen to adopt in btrfs in f34. 01:41:15 Michael: would have preferred the change in f34, but can do f33. Questions about public relations to answer. More testing would have helped there. 01:41:17 Matthias: it would still be useful to consider some of the ideas from Owen's proposal. 01:41:19 Owen: has concerns that the change is happening too quickly. Isn't going to push for f34 though. 01:41:21 - Recovery experience is a concern, particularly without the testing that we wanted (to have concrete numbers to point to). 01:41:23 - Wants to know how many people are testing Fedora with btrfs. 01:41:25 - We need a good story about performance. It's very subjective so ne need numbers and need to write the performance story so someone else doesn't. 01:41:27 Josef: 01:41:29 - Expanding staffing to work on recovery experience. 01:41:31 - Agrees with Owen's description of the performance challenge - going to assemble tests. 01:41:33 Owen: wants to avoid people being stuck in dracut with a prompt and no guidance. We need to get out of that situation. It's the "doesn't boot" scenario. 01:41:35 - Josef: we can prepare for this but it requires systemd cooperation and a few rounds of testing and iteration. 01:41:37 - Matthias: can we test for this? Neal: yes, if we can have a programmatic way of demonstrating recovery, we can simulate the experience. Davide: there is a kernel flag for making devices read only. There was a thread on this. 01:41:39 Allan: who is managing this effort (in a product management sense - all aspects)? 01:41:41 - Neal: there's a tracking ticket 01:41:43 - Matthias: the change proposal process assumes that the wiki page is the main place 01:41:45 - What about non-technical issues? Let's use the Taiga instance. 01:41:47 Neal: do we want to vote on the proposal? 01:41:49 - Allan: that ship seems to have sailed. 01:41:51 - Matthias agrees 01:41:53 - Langdon: the FESCo vote was just an engineering decision, and we discussed workstation making its own decision afterwards 01:41:55 - Owen: the proposal that went to FESCo wasn't pitched in that way; it's going to be too confusing if workstation goes its own way 01:41:57 - Langdon: there are still some questions in this group as to whether btrfs should be adopted; we ought to allow those people to express their view 01:41:59 - Jens: whether we approve for f33 or f34 is a potential complication in the vote 01:42:01 - Chris: we should trust the existing Fedora development process to iron out any issues 01:42:03 - Owen: that only applies if release blockers are technical bugs. We have other requirements, like there being documentation for read-only filesystem recovery. We need to think about the criteria that we care about. The blocker process is good for avoiding obvious defects, but does not catch additional requirements. 01:42:05 #agreed Plan is to proceed with btrfs by identifying potential issues/blockers and tracking them individually, rather than have a go/no go vote. No objections. 01:42:07 #topic default partitioning: solving the competition for free space 01:42:09 #link https://pagure.io/fedora-workstation/issue/152 01:42:11 #agreed Deferred. 01:42:13 #topic Guidelines for preinstalled and non-removable apps 01:42:15 #link https://pagure.io/fedora-workstation/issue/125 01:42:17 #agreed Deferred. 01:42:30 #endmeeting