[01:08] <callmepk> good morning
[01:27] <duflu> Morning callmepk 
[01:27] <callmepk> morning duflu 
[02:35] <pieq> hi duflu , callmepk !
[02:35] <duflu> Hey pieq 
[02:54] <callmepk> hi pieq 
[02:55] <pieq> So my desktop is still on 18.04. /home is on a different partition from /. What are the odds that I completely destroy everything if I install 20.04 and just replace / ? :D
[02:56] <pieq> (if I remember correctly, there is a "format" checkbox that is unticked by default, so if I point /home to my current home partition and leave this checkbox unticked, I should get all my files and configs back after the install, right?)
[03:03] <RAOF> pieq: You should keep your *user* configs and files (although backups are recommended, always!), but is there any particular reason not to just use `do-release-upgrade`? (You'll need to pass an option to make it consider 18.04→20.04 before we release 20.04.1)
[03:05] <pieq> RAOF, good question. A vague feeling of "it's cleaner this way". Also, last time I tried this (16.04 → 18.04), the laptop I tried it on miserably failed (a bad graphics driver issue I guess, I had no more X.org so it was very hard to recover)
[03:05] <pieq> RAOF, When is 20.04.1 due? Beginning of August?
[03:08] <RAOF> https://wiki.ubuntu.com/FocalFossa/ReleaseSchedule says “End of July”
[03:09] <pieq> RAOF, OK. I might just try the do-release-upgrade and see, then. It's a desktop with an AMD GPU and nothing too fancy (no wireless, no BT, ...), it shouldn't be too bad in terms of support
[03:10] <RAOF> My understanding is that `do-release-upgrade` is expected to *work* now, but we let the release see the first round of SRUs and soak for a bit.
[03:10] <RAOF> If you have any problems, please file bugs 😀
[03:11] <pieq> RAOF, I will, don't worry (I'm a QA engineer, filing bug is my passion! :D )
[03:11] <pieq> filing bug*s*
[03:12] <pieq> RAOF, I found some pulseaudio issues when upgrading another laptop from 19.10 to 20.04, but yeah, it was on a laptop, I don't expect to see this kind of issues on a desktop (but you never know! :))
[03:12] <pieq> RAOF, anyway, thanks for your answers!
[06:14] <didrocks> good morning
[06:17] <duflu> Hi didrocks 
[06:18] <didrocks> hey duflu 
[06:41] <oSoMoN> good morning desktoppers
[06:41] <oSoMoN> happy Monday!
[06:55] <duflu> Hi oSoMoN. Happy Monday
[06:55] <oSoMoN> hey duflu, how are you?
[06:58] <duflu> oSoMoN, going well. You?
[07:14] <oSoMoN> duflu, yeah, I'm good and very relaxed after a 4 day week-end
[07:23] <didrocks> salut oSoMoN !
[07:29] <jibel> hi all
[07:30] <didrocks> salut jibel 
[07:31] <jibel> salut didrocks, ça va bien?
[07:33] <oSoMoN> salut didrocks, jibel 
[07:34] <didrocks> jibel: ça va, et toi ?
[07:35] <jibel> didrocks, bien, toujours l'été ici, on en profite d'autant plus que les touristes ne sont pas là
[07:35] <jibel> natation et rando pour rattrapper le retard dû au confinement
[07:35] <didrocks> certes :)
[08:16] <duflu> Hi jibel 
[08:16] <jibel> good afternoon duflu 
[14:16] <Saviq> didrocks, jibel: hey folks, one thing just came to mind re: zsys - think it'd be possible to annotate the snapshots with what caused them? like what `apt` command was being run, stuff like that?
[14:20] <didrocks> Saviq: it’s an idea we had but we didn’t capture that on a bug report, mind doing so? (I’m afraid on a stable release that most of them will be "do-release-upgrade" and won’t really help though)
[14:25] <juliank> didrocks: I just setup a fully encrypted zfs laptop by hacking around the zsys ubiquity script, that was easy. need to play around with it more, but looking good
[14:25] <juliank> didrocks: The updating grub menu at apt install end is a bit slow
[14:26] <juliank> The installer experience was odd - it first formatted sda2 as / (ext4), and created a swapfile in it, before reformatting it as swap (sda3 was bpool, sda4 rpool)
[14:26] <juliank> Just to give you my impressions :)
[14:27] <juliank> (write speed seems to be about 60 MB/s in encrypted rpool, vs 100 MB/s in the unencrypted bpool; but they also have different recordsize [I increased recordsize to 512K in rpool])
[14:28] <didrocks> juliank: we need to test recordsize impact. The installer experience is unfortunately hackish, but normal users won’t see the trick ext4 then ZFS
[14:28] <jibel> juliank, we couldn't bypass partman, so we let it do it's thing then overwrite with our layout. That's weird indeed but didn't find a better way.
[14:28] <jibel> its*
[14:29] <didrocks> which is unfortunate due to how ubiquity does with partman
[14:29] <juliank> ack
[14:29] <didrocks> juliank: which encryption did you use?
[14:29] <juliank> didrocks: aes-256-gcm
[14:29] <didrocks> ack, the next default one :)
[14:29] <juliank> yeah
[14:29] <didrocks> just checking you didn’t shoot in your feet :)
[14:50] <Saviq> didrocks: no worries, will log :)
[14:51] <didrocks> thx!