gchristensen changed the topic of #nixos-dev to: NixOS Development (#nixos for questions) | https://hydra.nixos.org/jobset/nixos/trunk-combined https://channels.nix.gsc.io/graph.html | 18.09 release managers: vcunat and samueldr | https://logs.nix.samueldr.com/nixos-dev
Lisanna has quit [Quit: Lisanna]
Lisanna has joined #nixos-dev
Lisanna has quit [Client Quit]
Lisanna has joined #nixos-dev
lassulus_ has joined #nixos-dev
<ekleog> Is there any logic in the Build ID on hydra? Like, are them monotonically increasing or whatever?
<ekleog> (wrt. the commit version, not wrt. time of end of the build)
<ekleog> (unrelatedly… just noticed the most recent build appears to be on the left side of the chart, I've been wrong on that for like one year of looking at it \o/)
lassulus has quit [Ping timeout: 264 seconds]
lassulus_ is now known as lassulus
<clever> ekleog: every time an eval finishes, hydra will assign the jobs that differ from the previous eval a new sequential buildid#
<clever> ekleog: if it differs from the previous eval, but was built before, it gets a new id, but later finishes instantly as a cached build
<clever> and it shares the numbers across all builds
<ekleog> clever: ok thanks! :)
<ekleog> so it's sequential per end of eval and/or manual trigger, iiuc
<clever> ekleog: https://github.com/NixOS/hydra/blob/master/src/script/hydra-eval-jobset#L672-L677 is the rough location where that logic occurs
<ekleog> 'k thanks! :) (though I must say I don't speak Perl yet… should really get to learn it some day)
<ekleog> so now I think I more or less understand the basics of hydra's UI… I still can't see the reason why https://hydra.nixos.org/eval/1485074#tabs-now-fail made so many packages fail, like qtwebkit, while the two only commits compared to previous evaluation are “ruby: 2.5.2 -> 2.5.3” and “kitty: fix locating libstartup-notification-1” o.O
* ekleog thought it was him misreading the UI and taking things against the passage of time or something like that
<clever> ekleog: it might also be a random failure, checkout the exact rev listed under inputs and try building it locally
<ekleog> well it looks like a syntax error from the log, so random failure sounds unlikely, but going to try :)
<ekleog> (then, it's ike, ISTR it does have issues… if that's again the case maybe it'd be better to just disable it, and that'd actually make channels move faster?)
<clever> ekleog: where do you see the syntax error?
<clever> i see an eof in python code
<ekleog> yeah
<ekleog> can be transient failure due to bad RAM or so, I guess
* ekleog still downloading deps currently
<clever> it could also be the auto-gc stuff
<clever> ive got 13% of webkit built
<ekleog> you're going to have finished before I get deps downloaded :D
<clever> and i'm not even on my good internet today
<clever> 38%
<ekleog> still downloading :D
<clever> eek, it decided to fire up 7 gcc's in paallel, each eating nearly 2gig of ram
<ekleog> nice, hope you've got spare ram :D
<clever> it kept killing render processes in chrome
<clever> which has a nasty habbit of switching focus to the tab that just died
<clever> making it imposible to do anything because it keeps stealing focus :P
<clever> but i'm over that hump, and each gcc is using <500mb now
<ekleog> :D
<clever> make needs a better concept of resource usage
<clever> -j8 is fine 99% of the time, until it decides to run 8 things, each needing 2gig of ram, thats half the ram in the machine!
<clever> and thats why i say you should always have swap, even if it can harm performance
<clever> a little bit of slowdown is better then a hard fail
<clever> 79%!
<ekleog> still downloading \o/
<clever> 88%
<clever> 90%
<clever> ekleog: i think hydra failed at 83%?
<clever> yeah, it did
<clever> that build should probably be restarted then
<clever> ekleog: make install!
<ekleog> :D
<ekleog> hmm, now who can restart builds
<ekleog> hmm… randomly, andi-? can you restart https://hydra.nixos.org/build/83066260 ?
<ekleog> (also, I think maybe ike should be disabled / fixed? it appears to have been the source of every transient build failure I've heard of until now, maybe it's bad disk/ram? don't know whom to poke about this, though)
sir_guy_carleton has quit [Quit: WeeChat 2.2]
<samueldr> Ike again?
<clever> samueldr: yep
<clever> ive confirmed the webkit builds locally
* clever heads off to bed
jtojnar has joined #nixos-dev
orivej has quit [Ping timeout: 268 seconds]
drakonis has quit [Quit: WeeChat 2.2]
<ekleog> (triage) suggest closing https://github.com/NixOS/nixpkgs/issues/18194
<{^_^}> #18194 (by elasticdog, 2 years ago, open): pythonPackages.gdal broken on x86_64-darwin
<andi-> ekleog: done
<ekleog> thanks! :)
<andi-> that reminds me of that 1-line single-build-restart change I have somewhere for hydra.
<ekleog> (btw, there should be a list somewhere of people who can restart hydra jobs so we could just randomly pick someone, but just like for removing ike I don't know who to contact to actually make this change :/)
<andi-> Should open a PR
drakonis has joined #nixos-dev
drakonis_ has quit [Ping timeout: 252 seconds]
orivej has joined #nixos-dev
orivej has quit [Ping timeout: 245 seconds]
lopsided98 has quit [Ping timeout: 260 seconds]
lopsided98 has joined #nixos-dev
phreedom has quit [Ping timeout: 256 seconds]
phreedom has joined #nixos-dev
xeji has joined #nixos-dev
orivej has joined #nixos-dev
orivej has quit [Ping timeout: 252 seconds]
orivej has joined #nixos-dev
vcunat has joined #nixos-dev
<ekleog> andi-: but… to where? :/
xeji has quit [Ping timeout: 252 seconds]
<andi-> andi-: hu? I meant the Hydra change for restarting a single build/job
<ekleog> oh, sorry ^^' I just came back here without the message timing in mind and thought you were answering me to make a PR removing ike…
<andi-> ekleog: there seems to be a repo (nixos-org-hardware?) Where all machines are listed... But on the other hand there seems to be a deadlock issue which was attempted to be fixed. Not sure if the fix was ever applied there..
sir_guy_carleton has joined #nixos-dev
<ekleog> nixos-org-hardware? can't find anything that sounds like it :/
<andi-> Will check when I am at the computer later on.. Just about to board.
<ekleog> no problem, there's no emergency :)
<ekleog> makefu: That's the repo for hardware-configuration.nix, isn't it? I'm looking for the repo that lists hydra builders to remove one
<makefu> mhhhh damn, no idea sorry
<ekleog> meh, guess it's maybe not even public
<ekleog> nixos is in dire need of a good list of its members with who has access to what
<ekleog> but, like, I don't even know who to poke to get this list :D
<sphalerite> ekleog: https://github.com/orgs/nixos/teams but it's only visible to members of the nixos org >.<
<ekleog> sphalerite: yeah, that's exactly the problem :p also, I guess there isn't who has rights on hydra, the website, etc. in it?
<sphalerite> yeah true
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos-dev
<jtojnar> bah. the rpath stuff is killing me
<jtojnar> nothing really changed on meson side IIRC so I do not understand why it is suddenly an issue
<LnL> they changed some stuff around for darwin, but I did some testing and that fixed the install name issues we had before
<LnL> maybe that influenced the rest somehow?
drakonis_ has joined #nixos-dev
drakonis has quit [Ping timeout: 264 seconds]
genesis has quit [Quit: Leaving]
genesis has joined #nixos-dev
genesis has quit [Remote host closed the connection]
xeji has joined #nixos-dev
drakonis has joined #nixos-dev
sir_guy_carleton has quit [Quit: WeeChat 2.2]
orivej has quit [Ping timeout: 240 seconds]
vcunat has quit [Quit: Leaving.]
<nbp> Do we have an ARM64 NixOS yet?
<simpson> I think so, but I don't think that we have any prebuilt images for boards.
<sphalerite> nbp: we have aarch64 netboot images iirc
<sphalerite> nbp: in any case there are two aarch64 machines building for hydra
<sphalerite> and the community box also running nixos
<nbp> sphalerite: ok, so there is some hope for have pre-build binaries too :)
<sphalerite> nbp: absolutely, the coverage is pretty great.
<sphalerite> nbp: I run nixos on my aarch64 chromebook more or less without a hitch :)
<nbp> sphalerite: is it good enough for building applications, such as firefox?
<sphalerite> nbp: firefox runs fine
<nbp> sphalerite: hum … interesting.
<nbp> sphalerite: Thanks for the input :)
<sphalerite> built by the big machines in hydra of course
<sphalerite> I wouldn't want to build it on the chromebook :')
<sphalerite> I'll be bringing it along to nixcon as well if you want to see
genesis has joined #nixos-dev
orivej has joined #nixos-dev
orivej has quit [Ping timeout: 240 seconds]
orivej has joined #nixos-dev
orivej_ has joined #nixos-dev
orivej has quit [Ping timeout: 268 seconds]
drakonis has quit [Quit: WeeChat 2.2]
drakonis has joined #nixos-dev
aanderse has joined #nixos-dev
cransom has quit [Quit: WeeChat 2.2]
cransom has joined #nixos-dev
<thoughtpolice> https://github.com/NixOS/nixpkgs/pull/48919 for anyone interested in an easy PR
<{^_^}> #48919 (by thoughtpolice, 5 minutes ago, open): nixos: fix #48917 by setting SYSTEMD_TIMEDATED_NTP_SERVICES
<thoughtpolice> ekleog: The nixos.org configurations are in https://github.com/NixOS/nixos-org-configurations, which contains some traces of the hardware used for the build machines, though it doesn't seem comprehensive
<thoughtpolice> However, these don't include things like the info on the packet.net machines for nix.distributedBuildMachines (or whatever it's called), so I don't think that's complete
<thoughtpolice> There's also https://github.com/NixOS/nixos-org-configurations/blob/251de3f83b077bc90dcaf2e75b1b0556a6324ace/delft/provisioner.nix#L40 which indicates that there's a secondary custom 'machines' file which lays out the builders. So perhaps they are added manually through that
drakonis has quit [Quit: WeeChat 2.2]
drakonis has joined #nixos-dev
drakonis_ has quit [Ping timeout: 246 seconds]
orivej_ has quit [Ping timeout: 264 seconds]
drakonis has quit [Ping timeout: 252 seconds]
orivej has joined #nixos-dev
drakonis has joined #nixos-dev
sphalerit has joined #nixos-dev
xeji has quit [Quit: WeeChat 2.2]