<notgne2>
adisbladis: hey, I just opened a PR up for pnpm2nix as again it had some issues on some of my projects, it's a bit of a mess but at least documents what I needed to do to make it work https://github.com/adisbladis/pnpm2nix/pull/15
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/JveWy
o1lo01ol_ has quit [Ping timeout: 260 seconds]
orivej has joined #nixos
turona has quit [Ping timeout: 246 seconds]
turona has joined #nixos
<{^_^}>
[nixpkgs] @artemist opened pull request #77288 → factorio: update all versions to 0.17.79 → https://git.io/JveW9
o1lo01ol1o has quit [Ping timeout: 265 seconds]
<adisbladis>
notgne2: Great :)
<adisbladis>
notgne2: Btw, do you want commit bit? You're doing good work.
fenedor has quit [Read error: Connection reset by peer]
<notgne2>
adisbladis: that would be nice, but most of my changes are the ones I am most unsure of, so I think I would end up making a lot of PRs either way
anderslundstedt has quit [Ping timeout: 268 seconds]
anderslundstedt has joined #nixos
jluttine has quit [Ping timeout: 265 seconds]
vidbina has quit [Ping timeout: 258 seconds]
vld has joined #nixos
CMCDragonkai has joined #nixos
<CMCDragonkai>
If an application inside the container requires CA certs, is it sufficient to just add `cacert` to the `buildInputs` of the derivation? I'm using `dockerTools.buildImage`
Ralith_ has quit [Remote host closed the connection]
<emptyflask>
is anyone familiar with elasticsearch on nixos? I opened a PR (https://github.com/NixOS/nixpkgs/pull/68120) months ago and it's just sat there... I'm wondering if there's some better way to deal with plugins
<emptyflask>
since there's a good number of official plugins, and only a few of them are listed in nixpkgs
inferencerules has quit [Killed (verne.freenode.net (Nickname regained by services))]
inferencerules has joined #nixos
inferenc1rules has joined #nixos
<inferencerules>
ryantm: thanks for the help with imports yesterday; currently it's causing an infinite loop somewhere but that's definitely progress over not having any effect at all
<{^_^}>
[nixpkgs] @worldofpeace opened pull request #77294 → nixos/systemd-lib: don't fail on systemd.packages duplicates → https://git.io/JveBv
zeta_0 has quit [Remote host closed the connection]
kenjis has joined #nixos
<evanjs>
clever: yeah.... trying to find refer{rers,ences} but not getting too far :D
<Guest59_>
Any way to get a nix-shell going with llvm9? configurePhase won't work with: nix-shell '<nixpkg>' -A llvmPackages_9.llvm
<evanjs>
the only file in the derivation is... manifest.nix which contains `[ ]`. Like what
<clever>
evanjs: thats an empty nix-env -i profile
<clever>
evanjs: it happens if you nix-env -e everything
<evanjs>
clever: any reason you can think of it'd stick around like that and block rebuilds and etc until I remove it?
<clever>
evanjs: need more context on what your doing
mexisme_ has joined #nixos
<evanjs>
clever: not entirely sure what triggers it, but I did find this in the nixos-upgrade journal "skipping corrupt system profile entry ‘/nix/var/nix/profiles/system-838-link’".
boxscape has quit [Remote host closed the connection]
<evanjs>
It comes up every few times I try to either rebuild, collect garbage, or nix-env -iA something. Not quite sure which :D
<clever>
evanjs: does the symlink point to something?
<inferencerules>
infinisil: much appreciated; I won't get to debug until tomorrow morning again due to network badness here (my 2020 resolution is to stop using peered connections) but I'll probably have to take you up on that if you're around, fellow-inf
<evanjs>
clever: there's nothing but this one right now system-838-link-2-link -> /nix/store/57904rflsfh46y7fxi62rzc7sidc9vmq-user-environment
<evanjs>
when it stops what I'm doing, there's typically a system-838-link-1-link and system-838
<clever>
evanjs: you shouldnt have n-link-n-link in the names
<clever>
evanjs: that implies your -profile is wrong?
<clever>
evanjs: check ~/.bash_history on each user, what is every command you ran with system-838 ?
<clever>
evanjs: it should be pointing to a profile under per-user
<evanjs>
*checks ~.nix-profile*
<evanjs>
Oh look... It's red...
<clever>
likely because you cleaned the profile up
<clever>
but your next `nix-env -i` will recreate the mess
<evanjs>
ah. that explains this other error from earlier.... error: opening lock file '/nix/store/4x0hnmnpx9q3l0kdmyhxq695v4p9y2ic-nixos-system-nixtoo-19.09.1664.968381812b4.lock': Read-only file system
<evanjs>
Now that makes total sense :D
<evanjs>
(that's the missing profile that's linked to ~/.nix-profile ^)
<clever>
evanjs: you probably wanted `nix-env -p /nix/var/nix/profiles/system --switch-generation 838
<evanjs>
clever: ah that makes sense. No wonder almost everything else worked fine and nixos-rebuild skipped over it
<{^_^}>
[nixpkgs] @jonringer pushed commit from @bcdarwin to master « pythonPackages.crc32c: init at 2.0 »: https://git.io/JveRj
<kapil_>
I want nixos manage everything
doublex_ has quit [Ping timeout: 260 seconds]
<inferencerules>
ryantm: appreciate it; I'll get the updates up on github tonight and drop back here
<inferencerules>
really need to wrap my head around the imports stuff for overlays and whatnot, I'd like to start contributing to the docs on the wiki and the manual and the current state of things is pretty confusing imo
srid has quit [Quit: Connection closed for inactivity]
<kapil_>
how to install 64 bit when i do nixos-install?
<simpson>
If you booted in 64-bit mode, then you'll get a 64-bit userland and kernel by default.
<kapil_>
thanks.
emptyflask has quit [Quit: WeeChat 2.6]
<kapil_>
I first time installed nixos successfully! (after many try)
<simpson>
Congratulations! Welcome to the party.
<kapil_>
thanks
<kapil_>
where is my nix configuration file after install?
<kapil_>
oh i get it at /etc/nixox/configuration.nix
<simpson>
It'll be at /etc/nixos/configuration.nix; it's in the same spot as when you mounted during install.
<jackdk>
yep
<kapil_>
i changed it again but when i run nixos-install, it says mount point /mnt doesnt exist
mexisme has quit [Ping timeout: 240 seconds]
<simpson>
When you're booted into your system, you'll want to run `nixos-rebuild switch` instead. That will build your new system configuration and try to switch to it.
<kapil_>
thanks a lot
lord| has joined #nixos
<ryantm>
kapil_: Welcome!
<kapil_>
thanks, i want to move from arch linux to nixos
<kapil_>
is everything is managed in `/etc/nixos/configuration.nix`?
<kapil_>
its so nice
<ryantm>
kapil_: Yes, but can eventually break it into other files if it gets too unwieldy.
<ryantm>
Those are both configuration options in your /etc/nixos/configuration.nix file.
<ryantm>
I mean, options you can put in there.
<ryantm>
The default configuration file has environment.systemPackages in there, but commented out, I think.
<kapil_>
thanks a lot
<ryantm>
It's a list of packages that have the exact same name as what you wrote after -iA in your nix-env command
<kapil_>
i am installing plazma desktop its taking too much time
<ryantm>
kapil_: If something is taking a log time, that might mean you configured something in a way that is not available in the NixOS.org binary cache, so nix tries to build it for you.
<ryantm>
That's usually a good sign, because it means you are getting stuff from the binary cache instead of building it.
<kapil_>
oh thanks.
<kapil_>
oh its copying mariadb-server i just tell him to install plazma desktop.
<kapil_>
i think its a dependancy!
<ryantm>
Yep, probably is.
<kapil_>
but it should show some progress bar
<ryantm>
There's a command called "nix why-depends" that shows why something is a dependency of something else. You could read about it in the manual https://nixos.org/nix/manual/
<ryantm>
Yeah, that would be nice to have a progress bar for nixos-rebuild switch.
endformationage has quit [Ping timeout: 260 seconds]
<ryantm>
Somewhat recently people added progress bars for `nix build`.
<kapil_>
thanks
lovesegfault has quit [Ping timeout: 258 seconds]
drakonis has quit [Read error: Connection reset by peer]
<kapil_>
can i have seperate /home partion for different user data?
<inferencerules>
what do you mean different user data?
<kapil_>
like /Downloads etc..
<kapil_>
on arch linux there is I have seperate partition for /home
<ryantm>
Sure, maybe look at /etc/nixos/hardware-configuration.nix to see what is already defined, the main option is "filesystems".
rogue_koder has quit [Quit: Konversation terminated!]
<ryantm>
kapil_: You generally need to package things like that if you want to use them in NixOS, which is the price you pay for purity and reproducibility. Luckily, we all can share the cost by contributing to nixpkgs.
<kapil_>
ryantm: oh i thank you for replying. it so much helpful
bhipple has quit [Remote host closed the connection]
<kapil_>
the above .nix file is very simple. it donesnt tell the system where to install libraries?
softinio[m] has joined #nixos
<ryantm>
kapil_: they'll get installed into /nix/store/*actor-framework*
<jackdk>
and those sorts of common details are abstracted over by the stdenv.mkDerivation function
<ryantm>
kapil_: Inside there, it will be laid out similarly to how you might expect the regular file hierarchy to go.
<kapil_>
oh great
<kapil_>
jackdk: thanks
<ryantm>
The mkDerivation builder will set PREFIX to the appropriate place $out, so make install will install it into there.
lorimer has quit [Read error: Connection reset by peer]
<ryantm>
or I guess cmake in this case.
lorimer has joined #nixos
<kapil_>
thanks. so i must have nixpkg for each package?
<ryantm>
kapil_: What are you trying to do?
<kapil_>
i have a c++ project. its a web server. it uses different libraries.
<kapil_>
some time i need to install some library from github
chagra has joined #nixos
<kapil_>
i can easily do it on arch linux
<ryantm>
I'm not super familiar with C++ development using nix, but if I were doing that, I would not install the actor-framework/caf package on my system, I would make a "default.nix" file in my C++ project and use nix-build. That default.nix file would specify cag/actor-framework as a dependency.
<ryantm>
It's probably possible to install them on the system and add them to some linker path too though.
<{^_^}>
{ available = <CODE>; description = "An open source implementation of the actor model in C++"; homepage = "http://actor-framework.org/"; license = <CODE>; maintainers = <CODE>; name = <CODE>; outputsT...
growpotk- has joined #nixos
<kapil_>
ryantm: thanks
ee194350 has joined #nixos
ee194350 has quit [Read error: Connection reset by peer]
ee194350 has joined #nixos
jb55 has quit [Ping timeout: 240 seconds]
ee194350 has quit [Ping timeout: 268 seconds]
cartwright has joined #nixos
cantstanya has quit [Ping timeout: 240 seconds]
ee194350 has joined #nixos
jb55 has joined #nixos
vld has joined #nixos
vld has quit [Ping timeout: 265 seconds]
hyper_ch2 has joined #nixos
xO1 has joined #nixos
CodeWarrior has quit [Ping timeout: 245 seconds]
CodeWarrior has joined #nixos
ee194350 has quit [Read error: Connection reset by peer]
<eoli3n>
i used kubernetes formation tool online which has a side panel for explaination and clickable commands and another panel with a term
Cale has quit [Ping timeout: 260 seconds]
<eoli3n>
i want to use a tool like that to make a prez at work
<eoli3n>
any idea ?
Chiliparrot has joined #nixos
Cale has joined #nixos
<inferencerules>
eoli3n: probably best asked in another channel, but do you mean you can click commands in the panel and see the execution in the terminal in the other panel?
<{^_^}>
[nixos-org-configurations] @zimbatm pushed commit from @turboMaCk to master « status-page: fix race condition in initialization. (#94) »: https://git.io/Jveuw
kenshinCH has quit [Ping timeout: 265 seconds]
<contrun[m]>
Mic92: nur-combined seems to be not updating for a while. The arc repo is quite outdated. I `curl -XPOST https://nur-update.herokuapp.com/update?repo=arc` yesterday. It's still not updated.
<Mic92>
you can ask him to remove this package or make it a git submodule
<Mic92>
Also evaluation might become non-reproducible
<arcnmx>
ah mm looks like it's complaining about ifd somewhere, and yeah, it might be nice to do that though it's under constant development
<Mic92>
I was thinking about having a standarized namespace within a repo where ifd is allowed
<Mic92>
So we would only load the attribute name and not evaluate the package itsel.
<arcnmx>
the ifd was just because it felt a bit weird to duplicate a derivation when it happened to also reside with the source but :(
<arcnmx>
I should try to add a CI check for that, though I expect there may also be some yarn2nix packages that are an ugly mess and may need idf too if it's evaluating absolutely everything...
smatting has joined #nixos
orivej has joined #nixos
peanutbutter144 has quit [Ping timeout: 265 seconds]
peanutbutter144 has joined #nixos
growpotk- has quit [Ping timeout: 258 seconds]
m0rphism has joined #nixos
o1lo01ol1o has joined #nixos
<evils>
Mic92: know anything about passthru.updateScript?
lsix has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
jerryaldrichiii has quit [Remote host closed the connection]
jerryaldrichiii has joined #nixos
o1lo01ol1o has joined #nixos
vld has joined #nixos
<{^_^}>
[nixpkgs] @basilgood opened pull request #77300 → rsyslog: include libmaxminddb → https://git.io/Jvezz
ng0 has joined #nixos
ng0 has joined #nixos
<evils>
manveru: i added passthru.updateScript to kicad, this reports as updating for kicad-5.1.5, but my script is mostly intended to update kicad-unstable (same package, but called with a different pname), is there a way to change this?
<kapil_>
i am installing qt by following above guide. do i need to write .envrc file?
<clever>
,libraries kapil_
<{^_^}>
kapil_: Don't install libraries through nix-env or systemPackages, use nix-shell instead. See https://nixos.wiki/wiki/FAQ/Libraries for details.
hyper_ch2 has quit [Remote host closed the connection]
<Avaq>
Hi folks! I've had this "error(s) occurred while switching to the new configuration" warning for some time now when issueing 'nixos-rebuild switch', and I was wondering if anyone knows what's going on, whether it's important, and how I can fix it.
<ar>
interesting. iwd for me tries to use resolvconf from systemd, which expects systemd-resolved to be running, instead of using the resolvconf from openresolv
<clever>
kapil_: its a short-term fix, the long-term fix is to figure out why dns isnt being set correctly
<ar>
(i also get no dns entries)
<ar>
kapil_: are you using iwd?
kenjis has quit [Ping timeout: 248 seconds]
<kapil_>
ar: i just fresh install
<kapil_>
i am using hetzner server
detran has quit [Ping timeout: 260 seconds]
<clever>
kapil_: does hetzner provide dhcp or require a static ip config?
<clever>
kapil_: qtcreator has some trouble on nixos, qt is a lot more manageable in a shell
dansho has quit [Quit: Leaving]
cfricke has joined #nixos
<kapil_>
thanks
eoli3n_ has quit [Quit: WeeChat 2.7]
eoli3n_ has joined #nixos
<boolman>
kapil_: I have the same problem with /etc/resolv.conf since I last updated ( im using the unstable channel ). did you find any solution to this?
<evanjs>
Talking with a coworker on vsphere support (e.g. https://github.com/NixOS/nixops/issues/727) anybody have an idea if this is something in nixops' future, or if there's any motivation for such support?
<{^_^}>
[nixpkgs] @adisbladis pushed to release-19.09 « python: Add support for installing Python eggs »: https://git.io/Jve6Y
<{^_^}>
nixops#727 (by techtangents, 2 years ago, open): vmware vsphere support
captn3m0- has quit [Ping timeout: 268 seconds]
ddellacosta has joined #nixos
wpcarro has quit [Ping timeout: 260 seconds]
<evanjs>
I'm seeing mentions of terraform or etc as a bridge between nixops and vsphere, was curious if there were any more direct options
<{^_^}>
[nixpkgs] @vcunat pushed to staging « gdk-pixbuf: upstream patch to fix the nixos test »: https://git.io/Jve6g
<{^_^}>
[nixpkgs] @aanderse opened pull request #77326 → nixos/httpd: symlink apache configuration to /etc/httpd/httpd.conf for use in the apachectl command → https://git.io/JvePX
<{^_^}>
[nixpkgs] @bjornfor pushed commit from @squalus to master « keepassxc: fix darwin build »: https://git.io/JvePM
<Henson>
hi everyone, I'm trying to run the NixOS swiftclient and am getting the error "ModuleNotFoundError: No module named 'pkg_resources'". This happens both when I've installed it with nix-env and when I'm running it inside a nix-shell using python37.withPackages. This happens both on my Ubuntu and NixOS machines. Any suggestions?
<worldofpeace>
Henson: can you put the backtrace into a pasting service?
<worldofpeace>
It sounds like the package is missing a runtime dependency setuptools
<Henson>
I have a question, too. I noticed that the OpenSSH-7.9p1 package in nixos-19.09 is vulnerable to a couple of bugs that are fixed in the Debian version. I patched this myself in a github fork of nixpkgs but haven't submitted a pull request yet. I noticed that in nixos-unstable and nixpkgs-unstable there's not OpenSSH-8.1p1 which fixes these bugs. Should I wait for OpenSSH-8.1p1 to make its...
<Henson>
way into the nixos-19.09 channel, or submit a pull request for the patch for OpenSSH-7.9p1?
hyper_ch2 has quit [Remote host closed the connection]
<veleiro>
or how about, how is mesa handled in nixos configuration? I don't see any configuration outside of xserver, which doesn't apply to me as I'm using Sway
<clever>
veleiro: hardware.opengl.enable = true; handles most of it
sigmundv_ has quit [Ping timeout: 260 seconds]
erasmas has joined #nixos
<veleiro>
oh thats true
hyper_ch2 has joined #nixos
<the_pumpkin_man[>
It seems nobody is using Openbox ... :(
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/Jve1m
<veleiro>
so what's the difference between environment.systemPackages vs adding it to a program.*.extraPackages
<veleiro>
is loading packages by the latter packages only the program provides?
yobj has quit [Remote host closed the connection]
<evils>
i believe if you do program.*.enable = false, the extraPackages also get not loaded
<gchristensen>
program.something.extraPackages = [ hello ]; puts the hello package in the PATH for "something"
<erictapen>
veleiro: what is done with the content of program.*.extraPackages is up to the module author. In the case of program.sway.extraPackages it is just added to env.systemPackages, so there is no difference.
bvdw has quit [Read error: Connection reset by peer]
bvdw has joined #nixos
<rycee>
Hmm, anybody seen `nix: src/libexpr/attr-set.hh:54: void nix::Bindings::push_back(const nix::Attr&): Assertion `size_ < capacity_' failed.`? I started getting it for pretty much any use of the `nix` command. Like `nix run nixpkgs.hello`, for example.
opthomasprime has quit [Remote host closed the connection]
cosimone has quit [Quit: Terminated!]
shibboleth has quit [Remote host closed the connection]
<Thra11>
What if nixos-rebuild asked for a "commit message", so that each generation would have a short message describing it instead of just a number and date?
hyper_ch277 has quit [Remote host closed the connection]
<gchristensen>
I did that for a while, Thra11, it was really rough for me especially while iterating on changes. maybe if `nixos-rebuild test` didn't take one, but `switch|boot` did
<Mic92>
gchristensen: my tool is quite simple compared to your ofborg infrastructure.
<Thra11>
gchristensen: Hmmm. How did you do it? I was thinking it should be optional (and maybe autofill the message for things like --upgrade)
monsieurp has quit [Remote host closed the connection]
<gchristensen>
Thra11: really, I made it so I couldn't rebuild my system without first commiting my nixos-config repository. I made my system depend on a "package" whose source was the nixos-config repository on my disk, and if git said the repo was dirty, failed
andymandias_ has quit [Ping timeout: 268 seconds]
<gchristensen>
Mic92: meh, it is all about how it is for the end user :)
<Thra11>
If you made the descriptions mutable, you could "squash" generations (keep the most recent of a bunch of small tweaks, but give it the proper message, and delete the others)
andymandias has joined #nixos
monsieurp has joined #nixos
<LinuxHackerman>
I've been meaning to wrap nixos-rebuild (or replace it) with a tool that will automatically add everything and commit the result, then merge it (with the "theirs" strategy) to a "live" branch or similar with the generation number in the commit message.
gxt has quit [Ping timeout: 240 seconds]
<LinuxHackerman>
That way I have a record of which exact config resulted in the system generation, and how it differs from the manually committed state.
<LinuxHackerman>
I think that would only work well if I only ever build on a single branch though, i.e. no feature branches or collaboration
wildtrees has joined #nixos
<niksnut>
nixos-rebuild --flake records the commit hash of the top-level repo, which allows the entire configuration to be reproduced
<niksnut>
it doesn't auto-commit dirty trees yet though
andymandias has quit [Ping timeout: 240 seconds]
<maurer>
gchristensen: What about just making the commit message optional, but present on all generations? It'd just be an empty string if you didn't pass one in
<maurer>
that way users who want it can have it, but you don't need it when iterating
<niksnut>
rycee: I think that's fixed in nix 2.3.2
<rycee>
niksnut: Ah, nice.
Blessjah has quit [Remote host closed the connection]
<rycee>
niksnut: Hmm, release-19.09 still has 2.3.1 from what I can tell, perhaps should cherry-pick the 2.3.2 commit?
<{^_^}>
[nixpkgs] @andir opened pull request #77331 → firefox: 71.0 -> 72.0 (and other firefox packages as well) → https://git.io/JveS2
vidbina has joined #nixos
cosimone has joined #nixos
opthomasprime has joined #nixos
opthomasprime has left #nixos [#nixos]
<{^_^}>
[nixpkgs] @jonringer opened pull request #77332 → cookiecutter: use python3 → https://git.io/JveSK
timstott has joined #nixos
viric has joined #nixos
<timstott>
Hi all and Happy New Year. I've been trying to install a package via nix-env and specify a function argument without success (can't figure out what I'm doing wrong). Provided the polybar derivation exposes a i3Support argument, shouldn't the following override the value? nix-env -iA nixos.polybar --arg i3Support true
<viric>
helllo
CMCDragonkai has quit [Quit: Connection closed for inactivity]
<viric>
mpv is set to have 'tv' and 'v4l2' in linux but the log says they are not found.
<viric>
I'll have a look
<infinisil>
timstott: I suggest an overlay like `self: super: { polybar = super.polybar.override { i3Support = true; }; }` in ~/.config/nixpkgs/overlays/polybari3.nix
mbrgm has joined #nixos
<infinisil>
timstott: Then `nix-env -iA nixos.polybar` will install that version
<infinisil>
Also subsequent polybar updates (with nix-env -u) will maintain this override
<infinisil>
Or alternatively I'd recommend a declarative installation method
<infinisil>
,declarative timstott
<{^_^}>
timstott: There are multiple ways of managing declarative profiles. 1) Attrset, compatible with imperative use of nix-env https://git.io/fAQHW ; 2) buildEnv, providing more control over the paths that are linked into the profile https://git.io/fp0aU ; 3) home-manager, providing nixos-like config for your ~ https://github.com/rycee/home-manager
<infinisil>
viric: In any case, if you need tv/v4l support, the comment I linked to mentions how you can get it in recent versions, since tv:// doesn't work anymore
veleiro2 has quit [Remote host closed the connection]
<viric>
infinisil: hydra says: TV Interface : disabled
<viric>
infinisil: getting the same hash, built in my computer it says TV Interface: yes
<{^_^}>
[nixpkgs] @andir pushed 10 commits to release-19.09: https://git.io/Jve9B
<viric>
I touch nixpkgs once every few months, apparently with same goal.
dejanr has joined #nixos
<viric>
19.09 mpv still doesn't have v4l support. grmbl
<infinisil>
Did you try av:// ..
<viric>
infinisil: I don't think it will work with that. Checking for libv4l2 support : tv-v4l2 not found
<infinisil>
Just try i!
<viric>
^ hydra
<infinisil>
it*
drgbr has quit [Ping timeout: 240 seconds]
drgbr has joined #nixos
<infinisil>
The comment I linked to mentions the support is coming from ffmpeg
<viric>
infinisil: OH it works.
<viric>
aaaaaaaaah ffmpeg. damn it. libav.
<viric>
so clever. Less work for mpv then
<viric>
Now can play again with the endoscope
siiky has joined #nixos
bbarker has joined #nixos
nore has quit [Ping timeout: 268 seconds]
<o1lo01ol1o>
Hydra is stuck evaluating things forever. I've kicked `hydra-queue-runner`, `hydra-evaluator`, cleared vcs cache, rebooted, bumped nixpkgs to something other than 19.03 (fails for other reasons) and moved it back, but it's still stuck. Is there a local state database I can rm?
<Ankhers>
Does anyone use home manager and the mozilla firefox overlay? If so, do you use programs.firefox in order to install it, or do you just specify the package in home.packages?
<__monty__>
zyeri: Yes, something about synthetic.conf.
Tucky has joined #nixos
<zyeri>
__monty__: I saw something related to that in a GH issue. Is that going to be the best way to install nix then? I don't want to accidentally break something important.
<__monty__>
Yep, afaik that's going to be incorporated into the nix installer.
<zyeri>
ok
<zyeri>
I'll have a look at doing that then. Thanks!
<__monty__>
newbienixuser: Hmm, I'm not sure what this "nix environment" is. $out is usually made available in the various "phases" of a derivation. Also available in nix-shells.
mexisme has quit [Quit: WeeChat 2.7]
<newbienixuser>
I guess I have a $out defined, but it's empty
<__monty__>
The variable or the file? Cause the latter is normal. That echo is to create some output.
newbienixuser has quit [Remote host closed the connection]
jgeerds has joined #nixos
kuznero has joined #nixos
<kuznero>
Hi All!
<qyliss>
,ask kuznero
<{^_^}>
kuznero: Just ask your question. It's the best way to know if anybody can help. Or look at other ways of getting in touch with the Nix community here: https://nixos.wiki/wiki/Get_In_Touch
ixxie has quit [Ping timeout: 260 seconds]
<kuznero>
:) my question is this: I enable local kubernetes master+node roles with easyCerts and masterAddress = localhost. But kubelet fails to start...
<kuznero>
Any hints of what might be going on?
knupfer has joined #nixos
<kuznero>
This is what I get from `journalctl -u kubelet`: Running with swap on is not supported, please disable swap! or set --fail-swap-on flag to false. /proc/swaps contained: ...
cosimone has joined #nixos
mbrgm_ has joined #nixos
<kuznero>
I should probably try to disable swap ;)
mbrgm has quit [Ping timeout: 246 seconds]
mbrgm_ is now known as mbrgm
turion has quit [Ping timeout: 260 seconds]
<gchristensen>
yep! disabling swap is the best route
<kuznero>
:+1: trying out, thanks
<kuznero>
Will it work the other way round? services.kubernetes.kubelet.extraOpts = "--fail-swap-on=false";?
<kuznero>
I mean disabling swap for developer environment is not super-duper critical, right?
<gchristensen>
that might work
<gchristensen>
I don't think I've ever been happy that my kernel swapped, so I'm partial to turning off swap, but yeah that might work :)
<kuznero>
:) trying it now
<simpson>
Hi! Stupid question about contributing to nixpkgs: Is Dhall source code allowed in nixpkgs?
<gchristensen>
I doubt it, simpson
<gchristensen>
what for?
<gustavderdrache>
kuznero: the --fail-swap-on=false flag is deprecated in later k8s versions
dashkal has quit [Changing host]
dashkal has joined #nixos
jonten has joined #nixos
<simpson>
gchristensen: Argument in a private channel. I'm curious and don't mind being wrong or right.
<kuznero>
gustavderdrache, which means I will then still need to disable it. Ok, disabling swap instead...
<kuznero>
Thanks
<qyliss>
Dhall would require IFD, wouldn't it?
<gchristensen>
simpson: I just can't really imagine what it would be doing
<wildtrees>
how do I see what packages are installed via nix, using nix-env ?
<gustavderdrache>
kuznero: looks like it's only deprecated in 1.17, so you'll be safe to use that flag instead for a few versions :)
<simpson>
gchristensen: It is no business of mine why anybody would write any code for nixpkgs. Presumably they'd be packaging something.
<gchristensen>
hehe
<gchristensen>
I mean there is no rule against it, I just can't imagine why it would end up in nixpkgs
<simpson>
Because somebody merged a PR, probably.
<rycee>
niksnut: The nix 2.3.2 cherry-pick fixed the issue for me :-)
<kuznero>
gustavderdrache, but 1.17 is already release and I suspect it will be updated soonish.... so, better be prepared. On that note... for some reason I cannot use `services.kubernetes.package = unstable.kubernetes` to use the package from the unstable channel which is currently 1.16.4. How do I reference newer package properly?
<__monty__>
gchristensen: As context for simpson's question. He's talking about using dhall-nix's integration with nixpkgs to write entire expressions in dhall using just a `dhallToNix` skeleton expression to facilitate this: https://github.com/dhall-lang/dhall-haskell/tree/master/dhall-nix
<gchristensen>
that sort of generated code is not something we'd like to have in Nixpkgs
<dminuoso_>
wildtrees: nix-env --query
<wildtrees>
thx dminuoso_ , just kinda found it in the man pages, the manual didn't seem to mention it :(
<dminuoso_>
wildtrees: Admittedly discoverability is bad. If you have any suggestions how to improve it...
<kuznero>
Hm... how do I properly disable swap in NixOS?
silver_hook has joined #nixos
<dminuoso_>
kuznero: You can override swapDevices
<wildtrees>
dminuoso_, this is my second go at nix, just using it as a package manager on ubuntu on windows atm, instead of nixos on a constrained vm on my somewhat constrained laptop
<dminuoso_>
You likely have to use mkOverride/mkForce
shibboleth has joined #nixos
<dminuoso_>
wildtrees: Well nix-env is not the ideal way to manage packages.
<dminuoso_>
wildtrees: You will most likely find systemPackages in your configuration.nix to be more reliable
<niksnut>
rycee: great :-)
<infinisil>
kuznero: Or just don't define swapDevices at all, the default is to not have swap
zeta_0 has joined #nixos
chiefgoat has quit [Ping timeout: 268 seconds]
lord| has quit [Read error: Connection reset by peer]
<dminuoso_>
Swap is overrated anyway.
<dminuoso_>
RAM is cheap
<o1lo01ol1o>
hydra has a postgres database, how can I access / modify / recreate it ?
lord| has joined #nixos
<{^_^}>
[nixpkgs] @vcunat pushed 196 commits to staging-next: https://git.io/Jve53
<kuznero>
dminuoso_, infinisil thanks, trying it now
<kuznero>
Is it possible to override something that is defined in hardware-configuration.nix, but from configuration.nix? Just redefining swapDevices = [];?
<o1lo01ol1o>
well, there's nothing obvious to look at here by way non-terminating evaulations.
erictapen has quit [Ping timeout: 258 seconds]
<hyper_ch>
hmmm, I have a very weird problem: I can't boot into newly created generations created from nixos master
<o1lo01ol1o>
tilpner: seems like I have a few problems. The first is that the declaritive spec jobset returns: The .jobsets jobset must only have a single job named 'jobsets'. However, that jobset only has one attribute called "jobsets" (which is a spec.json containing all the jobsets)
erictapen has joined #nixos
<tilpner>
o1lo01ol1o: I only set that up once and never touched it again
<tilpner>
o1lo01ol1o: Launched things via web UI and used journalctl -eu (-fu) on hydra-evaluator and hydra-queue-runner
<o1lo01ol1o>
tilpner: yea, I don't know what's going on. I tried to make a jobset that ran closure compiler on some things and it never finished evaluating. Now no builds are sent (though the queue runner runs) and there's that error with .jobset . . .
<tilpner>
o1lo01ol1o: Paste all of it, I can have a quick look. I'll probably find nothing, and let you wait for someone else
<kuznero>
It seems that changing my hardware-configuration.nix to have swapDevices = []; does not make the swap block deactivated even after the restart. Am I missing something? I hope I shouldn't remove the whole swap partition...
<tilpner>
o1lo01ol1o: The hydra-evaluator.service log
chiefgoat has joined #nixos
<tilpner>
Not *everything*, just the parts you can tell are from an eval you triggered on purpose
<tilpner>
kuznero: nixos-rebuild switch didn't print any warnings/errors?
<kuznero>
nope
<tilpner>
kuznero: Use nixos-option to verify you set it correctly
<o1lo01ol1o>
tilpner: Every jobset is found to be "unchanged" in the log save one: hydra-evaluator[845]: evaluation of jobset ‘org-jobsets:.jobsets’ failed with exit code 1
<kuznero>
tilpner, `nixos-option swapDevices` shows it is empty = []. But when I do `lsblk` I still see it is activated with `[SWAP]` label against my partition
<o1lo01ol1o>
tilpner: the only detail on that: The .jobsets jobset must only have a single job named 'jobsets' at /nix/store/d23qlm9mpx0dbwh73wr9csaajrmrl7lr-hydra-2019-02-01/bin/.hydra-eval-jobset-wrapped line 636.
<kuznero>
And surprisingly /etc/fstab does not have a swap declaration... interesting...
kuznero has quit [Remote host closed the connection]
mbrgm_ is now known as mbrgm
kuznero has joined #nixos
knupfer has quit [Ping timeout: 245 seconds]
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.3]
<kuznero>
tilpner, gchristensen when I run `swapoff -a` it works, but even having swapDevices = [], it enables swap on startup. That is weird as if there are other things that make it be enabled... Maybe the way I installed nixos in the first place when creating symlink to /dev/disk/by-partition/swap ?
zeta_0 has joined #nixos
<hyper_ch>
really weird.... grub just can't find kernel to boot from for some reasons....
srid has quit [Quit: Connection closed for inactivity]
<tilpner>
kuznero: Just to be sure, free -th shows activated swap?
<tilpner>
o1lo01ol1o: I can't tell, sorry
chloekek has quit [Ping timeout: 258 seconds]
<kuznero>
tilpner, if I read it correctly then `free -th` has a line `Swap: 15Gi 0B 15Gi`
<kuznero>
which likely means it is enabled
<tilpner>
Yes, it looks that way
zeta_0 has quit [Quit: rcirc on GNU Emacs 26.3]
zeta_0 has joined #nixos
fenedor has joined #nixos
<tilpner>
kuznero: I absolutely agree that you should be able to deactive swap usage with swapDevices, but I'm curious why you want to disable swap permanently while keeping the partition?
<tilpner>
(There is systemd target "swap.target", might be interesting to query what triggered it)
<kuznero>
tilpner, the idea is that while I want to run kubernetes locally I want it to be disabled. But in case I want to stop doing that, I would like to re-enable it easily.
<tilpner>
kuznero: "Since swap is managed by systemd, it will be activated again on the next system startup. To disable the automatic activation of detected swap space permanently, run systemctl --type swap to find the responsible .swap unit and mask it."
<kuznero>
tilpner, interesting... unexplored realms of systemd... I see that it shows this: 1 loaded units listed. How do I mask it then?
<kuznero>
tilpner, your suggestions on how to permanently disable swap in NixOS without removing partition helped. Thanks a lot! I summed it up in a public github gist, please review if you can: https://gist.github.com/kuznero/bffb7f9a0b9829bf036be09b058e4322
<kuznero>
tilpner, thanks for the tip on manual... will submit a PR
jonten has quit [Ping timeout: 260 seconds]
<kuznero>
Here is what I see in kubelet logs now: network flannel/mynet: failed to set bridge addr: "docker0" already has an IP address different from 10.1.0.1/24
kuznero has quit [Remote host closed the connection]