<day|flip>
just wondering. is ther a faster way to make fb0 used nouveaufb. im using modesetting for drivers
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
IRCsum has joined #nixos
clacke_movim has left #nixos [#nixos]
IRCsome has quit [Read error: Connection reset by peer]
<{^_^}>
[nixpkgs] @thoughtpolice opened pull request #64276 → mimalloc: init at 1.0pre104_1125271 → https://git.io/fj6ny
IRCsum has quit [Remote host closed the connection]
freeman42x has quit [Ping timeout: 252 seconds]
IRCsum has joined #nixos
<hyper_ch>
In a nixos module option: is it not possbile to have a type = types.path but then also default = null; meaning either provide nothing or give a full filesystem path?
zacts has joined #nixos
<clever>
hyper_ch: there is a types.nullOr types.path
phreedom has quit [Quit: No Ping reply in 180 seconds.]
<hyper_ch>
thx
IRCsum has quit [Remote host closed the connection]
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Supersonic has quit [Ping timeout: 264 seconds]
IRCsum has quit [Remote host closed the connection]
clacke_movim has joined #nixos
IRCsum has joined #nixos
Myhlamaeus1 has quit [Remote host closed the connection]
Supersonic has joined #nixos
adamantium has joined #nixos
domogled has quit [Ping timeout: 246 seconds]
<typetetris>
`nix-env -iA nixos.haskellPackages.ghcWithPackages'(p: [p.cabal-install])'` stopped working for me with nix 2.2.2, it now interprets '(p: [p.cabal-install])' as literal part of an attribute key. How can I do that now?
shandogs has joined #nixos
kvda has joined #nixos
<Miyu-chan>
Hm. That is weird.
<Miyu-chan>
You can do it with -iE, I think
IRCsum has quit [Remote host closed the connection]
<allenj12>
anyone else have trouble installing steam? seems to have run into an issue with cp `cp: cannot create regular file '~/.local/share/Steam/bootstrap.tar.xz'`
<allenj12>
when I try to run it for the first time
IRCsum has joined #nixos
zacts has quit [Quit: WeeChat 2.4]
<Miyu-chan>
Yeah, you can just go with -iE
kvda has quit [Client Quit]
<typetetris>
With `-iE` I have to import nixpkgs first in the expression and then it ignores my `~/.config/nixpkgs/config.nix`.
<{^_^}>
[nixpkgs] @jonringer opened pull request #64279 → pythonPackages.seekpath: mark as broken → https://git.io/fj6cK
<typetetris>
Miyu-chan, it loads my config now.
<Miyu-chan>
:)
<allenj12>
hmm, is it just me, or should more things be configurable through configuration.nix? for example is it weird to want to move the status bar in xfce around etc.. in configuration.nix?
<Miyu-chan>
allenj12: That's the scope of home-manager TBH.
vmandela has quit [Quit: Leaving]
<allenj12>
ahh good point!
<allenj12>
ive read a little about that, but havent tried it yet
<Miyu-chan>
If I had to describe it, it's NixOS, but for users.
<Miyu-chan>
I personally haven't tried it out myself.
<Ralith>
it is good
<Miyu-chan>
I think alternatively, less things should be configurable through configuration.nix.
<Miyu-chan>
configuration.nix should be more on the sysadminny aspect, and it actually is for the most part.
<Miyu-chan>
I think the only thing that directly affects users is environment.systemPackages.
<Miyu-chan>
Oh, and user environment variables.
<Miyu-chan>
I kinda want to move some of my stuff to home-manager, but I've already had too much stuff linked to configuration.nix.
<allenj12>
idk, could you not argue that the configuration should be able to reproduce a system completely assuming similar hardware? since nixos is in the devops realm
<Miyu-chan>
Yeah, that's one reason why I still use configuration.nix, because it's much easier to manage one file than two.
kvda has joined #nixos
<allenj12>
yea I think for that reason I would like everything to be able to be reproducable in a nixconfig
<allenj12>
so... is there a place I can view home-manager configurations? like if I want to set xfce configurations
zacts has joined #nixos
<Ralith>
man home-configuration.nix
drakonis_ has quit [Ping timeout: 252 seconds]
<allenj12>
hmm kk, I have to relogor install brbf
allenj12 has quit [Remote host closed the connection]
growpotkin has left #nixos [#nixos]
allenj12 has joined #nixos
<allenj12>
o yea, has anyone else had a problem with shutting down/rebooting/ and logging out just halting for a very long time/forever
wfranzini has quit [Remote host closed the connection]
hyper_ch2 has joined #nixos
<hyper_ch2>
Hmmm, in a let statement I try to convert an integer to a string using: notify.intervalString = toString $cfgZED.notify.interval; but it keeps complaining. So, how do I properly do conversion? Google was no help
hyper_ch2 has quit [Remote host closed the connection]
hyper_ch2 has joined #nixos
drakonis_ has joined #nixos
mexisme_ has joined #nixos
m0rphism has joined #nixos
Drakonis has quit [Ping timeout: 252 seconds]
hyper_ch2 has quit [Remote host closed the connection]
kvda has joined #nixos
sb0 has quit [Quit: Leaving]
noudle has joined #nixos
hyper_ch2 has joined #nixos
cxx_newb has joined #nixos
<hyper_ch2>
ok, it seems I figured it out meanwhile :)
alp has joined #nixos
sushi_ has quit [Quit: Konversation terminated!]
<cxx_newb>
Hey guys. I am starting out with nixos and need to adjust my config coming from arch. Some scripts reference stuff from /usr/share of other tools which doesn't work anymore, as all that is now in the nix store. An actual example is that the fzf completion scripts are currently in /nix/store/6dn8jhgymvr9ws7piv4n2zvmcbwijj8v-fzf-0.17.5-bin/share/fzf
<cxx_newb>
which is a path that will change with update (is that correct?). How can I reliably reference stuff in the nix store across updates?
cxx_newb has quit [Remote host closed the connection]
<hyper_ch2>
cxx_newb: you could look in /run/current-system/.....
cxx_newb has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
<cxx_newb>
I just had a disconnect. In case anyone answered to my question if kindly ask to repeat the answer :)
allenj12 has quit [Ping timeout: 260 seconds]
hyper_ch2 has joined #nixos
knupfer has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
alp has quit [Ping timeout: 264 seconds]
palo1 has joined #nixos
palo has quit [Ping timeout: 246 seconds]
palo1 is now known as palo
cxx_newb has quit [Remote host closed the connection]
dansho_ has quit [Remote host closed the connection]
dansho_ has joined #nixos
sb0 has joined #nixos
hyper_ch2 has joined #nixos
kvda has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
kvda has joined #nixos
kvda has quit [Client Quit]
hyper_ch2 has quit [Remote host closed the connection]
mexisme_ has joined #nixos
noudle has quit [Ping timeout: 264 seconds]
hyper_ch2 has joined #nixos
dansho_ has quit [Ping timeout: 268 seconds]
noudle has joined #nixos
hio has joined #nixos
noudle has quit [Ping timeout: 248 seconds]
noudle has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
<Izorkin>
How to nedd access to enable build in darwin?
mexisme_ has quit [Ping timeout: 245 seconds]
markus1189 has joined #nixos
markus1189 is now known as predator117
mac10688 has quit [Ping timeout: 268 seconds]
predator117 is now known as markus1189
clacke_movim has left #nixos [#nixos]
Edes has joined #nixos
<teto>
trying to write a nixos test. I have seen very little of perl. I would like to save the output of a command and echo it in the shell later on. I do " my ($retcode, $output) = $client->execute("tshark -2 -R 'mptcp' -r test.pcap -Tfields -e frame.interface_id"); print $output;" but the display is empty :s can i capture stderr as well ?
<Edes>
Does anyone know how to use gnome extensions on NixOS?
<Edes>
whenever I install one it breaks and screams at me for not having packagekit
<Edes>
but I'm not entirely sure if that's required
lacesz has quit [Remote host closed the connection]
mexisme_ has joined #nixos
o1lo01ol1o has quit [Remote host closed the connection]
<teto>
tdeo: made a small comment, looks good otherwise
<hyper_ch2>
I'm trying to overwrite a default existing configuration file with a newly generated one but I cna't figure out how to do that. I thought I could use environment.etc."file".write but nixos complains about it: https://paste.simplylinux.ch/view/cbca1114#L571-L574
<teto>
hyper_ch2: what's the error ? you have to remove the file to let nix in control of it ?
<tdeo>
so `environment.etc."...".source="${pkgs.buildEnv {name="something";ignoreCollisions=true;paths=[packages.zfsUser, <write the second file here>]}}/etc/zfs/zed.d";
<hyper_ch2>
tdeo: you lost me :)
alex`` has joined #nixos
<tdeo>
yeah, typing it in an irc line is not great haha
<hyper_ch2>
ah, I think I get it
<hyper_ch2>
tdeo: it wasn't you... it's just me.... being an average computer user :)
<hyper_ch2>
tdeo: still no idea what to use <write the second file here> for
<tdeo>
i'm not exactly sure either immediately, but you need a function that will create a file in the right directory with that text
<tdeo>
i don't belive writeText does this, but i will test
<tdeo>
yeah, it doesn't...
<hyper_ch2>
the (default config) file already exists at "${packages.zfsUser}/etc/zfs/zed.d/zed.rc" and I just don't know how to alter it. I guess I have to do that during the postUnpack phase
phreedom has quit [Ping timeout: 260 seconds]
<hyper_ch2>
I still haven't figured out how module files actually write the config files to nix store
psyanticy has joined #nixos
<tdeo>
there's a function called writeText that takes two arguments, the name and contents of a file, and returns a nix store file, for example
jackdk has quit [Quit: Connection closed for inactivity]
<hyper_ch2>
but I must ensure that my custom file will finally be linked up as /etc/zfs/zed.d/zed.rc
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/fj680
<manveru>
i only need `python2`, and `python3`, otherwise this damn cmake fails :(
<manveru>
tried all kinds of -D flags, but it seems to just ignore me totally, so i straced, and it seems to pick the first `python` it finds on $PATH and checks its version
<manveru>
so i can make it fail depending on the order i put them in buildInputs :P
<LnL>
if you put them in a buildEnv priorities should handle it
<manveru>
the generation of the `python` executable seems pretty hardcoded in the python derivation, so don't think there's a way other than wrapping it
<rnhmjoj>
hyper_ch2: you can use `environment.etc` to symlink files in /etc or `system.activationScripts` to execute an arbitrary shell script to create a file anywhere
shandogs has quit [Remote host closed the connection]
cpc26 has quit []
shandogs has joined #nixos
iyzsong has joined #nixos
ilikeheaps has joined #nixos
<dminuoso>
Is there a point to using `install` over `cp` in derivations?
shandogs has quit [Client Quit]
<etu>
dminuoso: "install" has flags to specify mode of files, which can be nice to do in the same action
<etu>
So it's "cleaner" code
<dminuoso>
etu: Ah gotcha.
<betaboon>
dminuoso: mainly use install instead of cp as it can also create directories and set permissions. so instead of `mkdir -p ...`+`cp`+`chmod`+`chown` just one simple `install ...`
<Izorkin>
Mic92: please rechek PR netdata. Update service netdata
<{^_^}>
[nixpkgs] @peti pushed to master « python-bugzilla: fix broken package name »: https://git.io/fj6RW
<dminuoso>
Unrelatedly, I have a package that relies on some source with no stable download URL. What mechanism do I use to ask the user to prefetch manually?
<{^_^}>
[nixpkgs] @michaelpj opened pull request #64298 → asciidoctor: include the asciidoctor-epub3 gem → https://git.io/fj60T
<fgaz>
Also unrelated: idea: a "mutual review" discourse thread where one can post one of their prs and get it reviewed in exchange for reviewing another one
<fgaz>
would it encourage bad behaviour (ie rushed reviews)?
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/fj60V
<rnhmjoj>
hyper_ch2: does the file already exist?
<hyper_ch2>
rnhmjoj: yes, it exists already... I tried to delete it in the post installation phase but that just gave me an error with symlink then
<hyper_ch2>
I just can't figure out how nixos modules actually write config files
ericsagnes has joined #nixos
rprospero has quit [Remote host closed the connection]
<rnhmjoj>
wait, are there two modules trying to link different files to the same path?
alex`` has quit [Ping timeout: 245 seconds]
asheshambasta_m has quit [Quit: Leaving.]
gyroninja___ has quit [Ping timeout: 246 seconds]
rprospero has joined #nixos
alex`` has joined #nixos
xcthulhu has joined #nixos
mexisme_ has quit [Ping timeout: 245 seconds]
<hyper_ch2>
rnhmjoj: I have no idea what I'm doing as nowhere there's some kind of explanation
<rnhmjoj>
hyper_ch2: the source of the error in nixos/modules/system/etc/make-etc.sh#33, when you create a symlink with the option environment.etc this script will be called to actually make the link. the script will refuse to create it if a symlink already exists and it's pointing somewhere else
<rnhmjoj>
if you deleted it and later reappeared i guess it means there's another nixos module making the link
rprije has quit [Ping timeout: 272 seconds]
<hyper_ch2>
so how to write the config file then to ${packages.zfsUser}/etc/zfs/zed.d/zed.rc ?
__monty__ has quit [Ping timeout: 258 seconds]
v88m has quit [Read error: Connection reset by peer]
v88m has joined #nixos
gyroninja___ has joined #nixos
<rnhmjoj>
is this the module? nixos/modules/tasks/filesystems/zfs.nix
<hyper_ch2>
rnhmjoj: yes
<hyper_ch2>
I want to add options for ZED
rprospero has quit [Remote host closed the connection]
<hyper_ch2>
I think I have it working now... but not sure
__monty__ has joined #nixos
<rnhmjoj>
ah, i see. one option is symlinking a whole directory
<hyper_ch2>
I know
<hyper_ch2>
I just thought I could easily just add a file
<rnhmjoj>
and that directory probably contains a different zed.rc, right?
noudle has quit []
<hyper_ch2>
maybe this helps as last thing in the let statement part pkgs.writeTextDir "etc/zfs/zed.d/zed.rc" zedConf;
<hyper_ch2>
rnhmjoj: yes, but I altered pkgs/os-specific/linux/zfs/default.nix to remove it
<hyper_ch2>
I tried with it being there... no luck... so I tried to remove it... no luck
xelxebar has quit [Remote host closed the connection]
<rnhmjoj>
ok, can't you avoid symlinking the whole directory? if there are only a few files you could create a link manually, and add zed.rc
xelxebar has joined #nixos
<hyper_ch2>
rnhmjoj: the symlilnking of the whole dir was already there :)
<rnhmjoj>
hyper_ch2: i think the problem is nix can't touch that directory becase it's inside the nix store and therefore read only. there probably is no other any besides creating the link individually or changing the zfs package.
Havvy has joined #nixos
rprospero has joined #nixos
Ralith_ has quit [Ping timeout: 258 seconds]
tput has quit [Quit: ERC (IRC client for Emacs 26.1)]
<hyper_ch2>
shouldn't that file be created somewhere?
gyroninja___ has joined #nixos
init_6 has quit []
Makaveli7 has quit [Quit: WeeChat 2.5]
<rnhmjoj>
hyper_ch2: is there no symlink in /etc/zfs/zed.d/zed.rc?
<rnhmjoj>
it seems ok now, i would expect it to work
Anymost has quit [Remote host closed the connection]
<hyper_ch2>
rnhmjoj: I just built, not rebooted
<hyper_ch2>
I did then find /nix/store -iname "zed.rc" -type f and it couldn't find my generated one... only the one that is there whit the system up and running now
hoijui has joined #nixos
<rnhmjoj>
hyper_ch2: it should be somewhere but i don't know if it's going to contain the name of the target (zed.rc)
hyper_ch2 has quit [Remote host closed the connection]
rprospero has joined #nixos
xcthulhu has quit [Quit: xcthulhu]
rprospero has quit [Client Quit]
<dminuoso>
Ohh wow Im silly.
hyper_ch2 has joined #nixos
hyper_ch2 has quit [Remote host closed the connection]
<dminuoso>
But Im still curious, why doesn't that work?
hyper_ch2 has joined #nixos
<hyper_ch2>
rnhmjoj: I got it /nix/store/11zyvpd6wrfk41zzw9by1x7gi81mn67p-etc-zed.rc
<hyper_ch2>
*"zed.rc"* helped
<rsa>
i've packaged a library where manually running unpackPhase, patchPhase, configurePhase and buildPhase builds it just fine, but when i try to nix-shell -p it it breaks at the configuring stage with "...setup: ./autogen.sh: /usr/bin/env: bad interpreter: No such file or directory"
<hyper_ch2>
well, find ..... *"zed.rc"* did the trick
<rnhmjoj>
ah, i wasn't familiar with -iname: i thought it was a variant that matched partially
<rnhmjoj>
it's just case insensitive, so you still need the wildcards
<hyper_ch2>
yes, I'm just used to use -iname and usually with grep I also use -i :)
<hyper_ch2>
file is generated fine... just one issue.... one option that should be commented out isn't :(
<hyper_ch2>
yeah.... there's the email addresses list attribute xxx = [ "xxx" "yyy" ]; and then I do check if cfgZED.email.addresses == null then %comment out% ..... but I guess the list attribute is never "null"
ma27 has joined #nixos
<rnhmjoj>
is the list going to be empty in that case? you could just check == []
<hyper_ch2>
ah that works? didn't know :) will test it
<michaelpj>
what's the general approach for packages with optional unfree dependencies? do something like have them nulled out by default so they can be overridden if someone wants the unfree version? gate on `config.allowUnfree`?
<hyper_ch2>
this generates the file: (if cfgZED.email.addresses == null then '' but this won't (if cfgZED.email.addresses == [] then ''
jackdk has quit [Quit: Connection closed for inactivity]
<hyper_ch2>
rnhmjoj: any idea why == [] won't work as anticipated?
iyzsong has quit [Ping timeout: 276 seconds]
orivej has quit [Ping timeout: 245 seconds]
detran_ has quit [Ping timeout: 272 seconds]
<dminuoso>
Mmm, systemd.services.<name>.description seems to be a misnomer - from what I can see its rather meant as an alternative name to pop up in systemd for the unit.
civodul has quit [Read error: Connection reset by peer]
<deech>
Hi, is it possible to write scripts in something other than Bash or does that invalidate all existing infrastructure wrt buildInputs etc?
hyper_ch2 has joined #nixos
ilmu has quit [Ping timeout: 252 seconds]
Guest68314 has joined #nixos
<hyper_ch2>
new irc client always gives me disconnect :(
<lordcirth>
deech, scripts for what purpose?
asheshambasta_m has joined #nixos
<tazjin>
has anyone here spent some time on the combination of Nix & Bazel? Getting things from Nix into a Bazel build is easy, the other way around not so much ..
hyper_ch2 has quit [Remote host closed the connection]
<Henson>
hello everyone, I want to use a programatically-generated list inside of a mkDerivation buildCommand string. I'm assuming I need to define the list outside of the buildCommand string and then refer to it inside? If so, how can I convert a list of strings to the string representation of it?
<clever>
> lib.concatStringsSep "\n" [ "a" "b" "c" ]
<{^_^}>
"a\nb\nc"
<clever>
> lib.concatMapStringsSep "\n" (n: "str is ${n}") [ "a" "b" "c" ]
<{^_^}>
"str is a\nstr is b\nstr is c"
Ralith_ has quit [Remote host closed the connection]
Ralith_ has joined #nixos
<clever>
Henson: ^^
mexisme_ has quit [Ping timeout: 272 seconds]
Drakonis has joined #nixos
orivej has joined #nixos
drakonis_ has quit [Ping timeout: 258 seconds]
lambda-11235 has joined #nixos
ambro718 has joined #nixos
<Henson>
clever: awesome, thank you
<ambro718>
I want to use ZFS with encryption and two mirrored disks. Since I also want swap/hibernation that rules out ZFS-level encryption. Swap should not be mirrored. So I am thinking cryptsetup directly on each disk, and partitions inside that, one disk has ZFS and swap inside and the other has just ZFS, and the two ZFS partitions are mirrored on ZFS level.
<ambro718>
But how can I ensure that the two disks have both the same cryptsetup passphrase, and that I only enter it once during boot?
<ambro718>
Another thing I want is to be able to boot off any disk in case one dies.
<Netsu>
clever: is there any way to fetch the directory locally. I use fetchurl with hash from 'nix-prefetch-url' but looks like it creates just empty one. On other hand 'requireFile' throw an error: https://gist.github.com/Pitometsu/c61a2e95fc97d618aa87eaaf08fe8f5f
<clever>
Netsu: the whole point of requireFile is to throw an error, that explains how to add the file with nix-prefetch-url or nix-store --add-fixed
<Henson>
sometimes when I run nixos-rebuild after changing the configuration file it needs to download certain missing dependencies like patchelf, gnumake, or other things. Do these tools belong to a certain suite that can be installed? Every time I garbage collect it has to re-download these things.
<ivan>
Henson: see if keep-outputs = true in nix.extraOptions keeps them around?
knupfer has joined #nixos
<ivan>
if they're still not retained maybe you can add them to systemPackages?
bgamari has quit [Ping timeout: 258 seconds]
<Henson>
things like gnumake-4.2.1 stdenv-linux patchelf-0.9
linarcx has joined #nixos
mara[m] has left #nixos ["User left"]
<ambro718>
Any idea why the google-chrome binary is 141 MB but the chromium binary is 231 MB ?
andreas303 has quit [Ping timeout: 260 seconds]
Tucky has quit [Quit: WeeChat 2.5]
<maurer>
The released chrome binaries have a lot of postprocessing done on them to keep binary size down
linarcx has quit [Client Quit]
knupfer has quit [Remote host closed the connection]
iqubic has joined #nixos
<maurer>
I think we strip them, but that's about it
knupfer has joined #nixos
andreas303 has joined #nixos
amakthel has joined #nixos
linarcx has joined #nixos
yuken has quit [Ping timeout: 248 seconds]
Mateon2 has joined #nixos
<iqubic>
I'm in deep trouble. I have somehow run out of space on my harddrive. I tried running nix-collect-garbage, but that failed because there wasn't enough space to set permissions on /nix/var/nix/per-user. What should I do?
<lordcirth>
iqubic, what filesystem?
<lordcirth>
normally only fancy filesystems like ZFS or Btrfs completely jam when full
Mateon1 has quit [Ping timeout: 272 seconds]
Mateon2 is now known as Mateon1
<lassulus>
nix-collect-garbage jams when the harddrive is full
<iqubic>
ZFS
<slabity>
iqubic: Do you have any temp directories that are on your drive? Sometimes /tmp or ~/.cache contain a ton of data
<lordcirth>
iqubic, ok. So try to find a file of at least a few MB that you can delete. Then "echo > file"
<lordcirth>
An rm requires a metadata allocation, but truncating it doesn't.
<LnL>
doesn't nix create a reservation file now?
<iqubic>
I more than 100 files were deleted from /tmp.
<lordcirth>
I'm not too familiar with NixOS, it might. I just know how to unjam a full ZFS.
<iqubic>
Here's what I'm actually going to do to save space: delete windows 10 from my machine. I haven't openned Windows in 6 months.
yuken has joined #nixos
<LnL>
this usually also helps nix-store --gc --max-freed 1m
<iqubic>
But that's a big task that I donåt want to deal with right now.
<LnL>
with increasing values so it's not one huge delete
<iqubic>
LnL: I still don't have enough space to do that.
<LnL>
yeah, with zfs you'll have to delete something first if you didn't set quota
<iqubic>
Setting permnssions fails.
<iqubic>
I deleted everything from /tmp and !/.cache.
<iqubic>
~/.cache I mean
<lordcirth>
iqubic, but do you have snapshots of it?
<lordcirth>
If you do, you won't actually free any space
<iqubic>
perhaps. Hol do I delete snapshots?
freeman42x has joined #nixos
bgamari has joined #nixos
<iqubic>
How do I delete ZFS snapshots?
yuken has left #nixos ["Leaving"]
zacts has joined #nixos
shabius_ has quit [Quit: Leaving]
<ambro718>
How would I achieve having mirrored boot partitions on two disks, so I can boot from the other disk of one fails?
iqubic has quit [Remote host closed the connection]
Jackneill has quit [Remote host closed the connection]
iqubic has joined #nixos
<iqubic>
So, what can I do here?
dm9 has quit [Quit: WeeChat 2.5]
<lordcirth>
iqubic, first, 'zfs list -t snap'
<lordcirth>
Then, if you have some which you wish to destroy, you can run 'zfs destroy <name>'
<{^_^}>
[nixpkgs] @eraserhd opened pull request #64310 → kak-ansi: init at 0.2.0 → https://git.io/fj6aw
<iqubic>
There are a ton of snapshots. How do I destroy them all?
shabius has joined #nixos
<lordcirth>
iqubic, well, first look for ones that are actually using much space, the "USED" column. Quite likely deleting the biggest one will be enough.
<iqubic>
I want to destroy all the snapshots.
<Ralith>
how does NixOS expose its certificate store to applications?
<makefu>
Ralith: SSL_CERT_FILE
<lordcirth>
iqubic, "zfs destroy -rnv storage@%" where "storage" is the name of the filesystem the snaps belong to. If the output looks fine, remove the -n.
<Ralith>
makefu: that doesn't seem to be set in a user environment, though?
<makefu>
correct, to get this i set environment.variables.SSL_CERT_FILE = "${pkgs.cacert}/etc/ssl/certs/ca-bundle.crt";
<iqubic>
Now I have enough space to delete stuff.
<makefu>
Ralith: but /etc/ssl/certs is also populated in nixos
<makefu>
so maybe some fallback mechanisms kick in
<lordcirth>
iqubic, excellent. Once all the garbage collection is done, you may want to examine your auto-snapshot settings and get it to delete really old ones.
<{^_^}>
[nixpkgs] @michaelpj opened pull request #64311 → mathoid: init at 0.7.2 → https://git.io/fj6aP
<Ralith>
ahah, thanks
<iqubic>
Lordcirth, I have to go now, but I will look into that.
<Henson>
ivan: thanks for the suggestion, I think keep-outputs is doing the trick
<kraem>
anyone running gnome-keyring? my keyring gives me a gui asking for my gpg pass but never when i try to ssh. SSH_AUTH_SOCK is not set on startup. any clues?
<Ralith>
kraem: I'd try using home-manager to configure it
judson_ has joined #nixos
<iqubic>
also, I'm going to need to mess with my partitions because I'm planning Windows10from my system, which is /dev/sda2 and /dev/sda3. I think. I'll be doing the
<iqubic>
checks before deleting a filesystem
<iqubic>
But that's a quest for another time.
judson has quit [Read error: Connection reset by peer]
iqubic has quit [Remote host closed the connection]
wildtrees has joined #nixos
alp has quit [Ping timeout: 252 seconds]
<mkaito>
what would a derivation look like that generates a valid profile? I was looking at Hail, and while I'm not planning to use it, it wants a jobset that generates a user profile which it then pulls. Other than imperatively calling nix-env, how?
<jD91mZM2>
So toPath isn't what one expects... How does one convert a string to a path in NixOS? `./. + string` misses a leading slash, `./. + ("/" + string)` works for relative paths, but not for absolute ones :|
bgamari_ has quit [Remote host closed the connection]
<jD91mZM2>
In Nix**
<iqubic>
Its 'een like 2 months since I last did this.
<mkaito>
slow hard drive?
<iqubic>
SsD
<mkaito>
huh
<mkaito>
well let me know how much it frees when its done
<iqubic>
14188 store path deleted 3902.06MiB saved.
<mkaito>
that's not even that much
<mkaito>
your ssd sucks :P
<iqubic>
Does it realuyw
<mkaito>
apparently
<iqubic>
*really?
bgamari has joined #nixos
<mkaito>
I don't think mine needs half an hour to collect 4 gigs
<iqubic>
Hard linking is saving *e 1.74.42 MiB
<iqubic>
Is that odd?
<mkaito>
not really. your drive is just somewhat slow.
<lordcirth>
iqubic, 1.74.42 ? what?
<iqubic>
1074.42 MiBs being saved by hard linking
<iqubic>
I hate thir keyroard.
<iqubic>
Using my phone to do irc, as I am on a train with no wifi
<mkaito>
yeah phone keyboards are satan
<clever>
jD91mZM2: / + "/foo"
<iqubic>
So, when my drive was full, I tried editing my configuration.nix. That didn't end well.
<iqubic>
My configuration.nix is now empty
<jD91mZM2>
clever: If you mean `/. + string`, that only works for absolute paths
<iqubic>
How do I load a ZFS snapshot?
<clever>
jD91mZM2: why are you trying to generate paths from strings?
<clever>
iqubic: what filesystem is the snapshot made from?
<lordcirth>
iqubic, didn't you just delete them all?
<jD91mZM2>
clever: I basically just want to normalize a path. I'm parsing the path from a TOML file because... reasons
<iqubic>
no I only deleted two that were 53 MB each.
<iqubic>
I still have the latest snapshots.
<clever>
iqubic: is the root fs dataset snapshots?
ambro718 has quit [Quit: Konversation terminated!]
<iqubic>
yes I have a snapshot of the file I want to recover.
<clever>
iqubic: ls -l /.zfs/snapshots/
<clever>
iqubic: oops, just /.zfs/snapshot/
<iqubic>
I have verified that the snapshot exists with ZFS list -t snapshot.
<clever>
iqubic: all snapshots are auto-mounted under /.zfs/snapshot/
<iqubic>
I don't have wifi right now, so I can't paste any command outplts now.
<jD91mZM2>
clever: Ok so for some reason I thought it would be a good idea to let a user of some nix expressions override the source using... TOML. I wanted to make it as easy as possible, but eh I can just use nix.
<clever>
iqubic: you can just poke around in /.zfs/snapshot/ and find your old configuration.nix
<iqubic>
Oh, so I can fetch the files I need?
<clever>
iqubic: yes
<lordcirth>
iqubic, yeah, just copy it out and put it back where you want ti
amakthel has quit [Ping timeout: 250 seconds]
<iqubic>
cool. glad I didn't delete all snapshots
<mkaito>
I reccommend git
<iqubic>
Also, maybe I should put my configs on github.
<clever>
mkaito: zfs snapshots are taken every 15mins automatically, so you dont have to obsessively make commits
<mkaito>
I didn't say anything about obsessions :D
<iqubic>
ZFS tells me I have some snapshots of /home, but /home/iqubic is empty in the snapshots
<iqubic>
What can I do?
v88m has quit [Read error: Connection reset by peer]
v88m has joined #nixos
<Henson>
is there a way to have multiple profiles put into your environment? For example I'd like to have the default profile, but another one with additional software overlaid together.
<iqubic>
'ecayse one important config file has 'een deleted, and I can't seem to recover it.
<iqubic>
*because
<iqubic>
I'm reallp really concerned.
andreas303 has quit [Remote host closed the connection]
Shoubit has quit [Ping timeout: 248 seconds]
iqubic has quit [Remote host closed the connection]
alp has joined #nixos
iqubic has joined #nixos
hoijui has quit [Ping timeout: 252 seconds]
<iqubic>
So I don't know what to do.
psyanticy has quit [Quit: Connection closed for inactivity]
<mkaito>
Henson: AFAIK you only have one profile active at any given point in time. You can switch to other profiles on the fly using nix-env though.
<Henson>
mkaito: ok, thank you
<mkaito>
maybe there's some buildEnv magic that can compose multiple profiles into one
<iqubic>
My entire WM config has vanish
<mkaito>
iqubic: on linux, running completely out of disk space is *really* bad.
<iqubic>
Because I did a foolish thing.
andreas303 has joined #nixos
<iqubic>
I know.
<mkaito>
it's very deep in do-not-want territory, sadly
<iqubic>
ZFS claims I have snapshots of /home, 'ut I can't seem to find those configs.
<mkaito>
I've never used ZFS, so no idea about that
<Henson>
iqubic: if zfs says you have snapshots, you could clone one of the snapshots and take a look at what's inside
<Henson>
iqubic: and then you'll be able to see what's in the snapshot. A clone is a RW copy of a snapshot.
<iqubic>
ah, I assume first arg is src, and second is destw
<clever>
Henson: no need to clone a snapshot to look at it
<clever>
mount -t zfs path/to/dataset@snapshot /mnt/
<Henson>
iqubic: just "man zfs" and read up on "zfs clone"
jb55 has quit [Ping timeout: 260 seconds]
<Henson>
clever: oh really? Is that a read-only mounting, then?
<clever>
yeah
<clever>
but if path/to/dataset is mounted at /path, you can also just look in /path/.zfs/snapshot/snapshot
<mkaito>
this is probably not a great time to ask... but how's ZFS on Linux these days? :P
<clever>
and it will auto-mount it (read only) for you
<Henson>
mkaito: I gave up on ZFS on linux a couple years ago and switch to BSD, and have never looked back.
<mkaito>
nice
<clever>
Henson: the .zfs directory is weird, is exists, but isnt visible in any directory listing
<Henson>
mkaito: but that doesn't answer your question, I guess
bakakuna has quit [Ping timeout: 248 seconds]
bakakuna1 has joined #nixos
<Henson>
clever: wow, I'm learning things about ZFS on FreeBSD in the #nixos channel :-)
<ejpcmac>
mkaito: It seems that ZFS on Linux is good enough for the FreeBSD project to switch to it instead of the current Illumos port.
domogled has quit [Ping timeout: 258 seconds]
<iqubic>
Clever, you said you said I could just look at /.zfs to find the snapshots. I found the snapshots. The three I looked at had a /home/iqubic folder, but all three were empty.
<clever>
iqubic: is /home/iqubic its own dataset in zfs?
<iqubic>
that is what concerns me.
<lordcirth>
iqubic, can you pastebin 'zfs list' please?
<clever>
iqubic: then you want /home/iqubic/.zfs/
<clever>
each filesystem has its own snapshots and its own .zfs
<iqubic>
Oh. Yeat, that would solve my issues.
<Henson>
clever: holy crap, I never knew about the .zfs directories. That's amazing!
<Henson>
wow, I'll never have to create a clone to look at a snapshot again.
bakakuna1 is now known as bakakuna
<clever>
Henson: even more fun, you can use mkdir in the snapshot dir, to create snapshots
<iqubic>
That solves my issfe Clever.
adamantium has joined #nixos
<iqubic>
*issue.
<iqubic>
All files have been restored.
<Henson>
clever: on my system it says operation not permitted. There's probably a zfs option somewhere that affects that.
<___laika>
what would people say is the most idiomatic way to pin nixpkgs in a project? i feel like i have seen mixed recommendations between using `nix-prefetch-git/url` in the command line to generate a json file, and having the url and hash in code to pass to `builtins.fetchTarball`, and i cant tell if there's a meaningful difference between the two
emptyflask has quit [Quit: WeeChat 2.4]
<manveru>
___laika: the json is useful if you're using it with other tools as well, or generate it via CI or something
<manveru>
i don't think there's a real convention...
<manveru>
i have a little script that just spits out the nix code for latest nixpkgs so i paste it in there...
<noonien>
hello folks
<noonien>
are there any examples i can look at for using imports in overlays to extend them?
<noonien>
i have one main overlay.nix, in which i want to import multiple others
<nervengift>
is there a good example of a package that supports plugins which it expects in a certain directory? I managed to write a derivation that builds the plugin (rofi-blezz) as a shared lib but obviously I can't install it to the original program's directory in the nix store...
<clever>
nervengift: generally, the package should accept an env var pointing to a dir of plugins, and then you use buildEnv to merge all the plugins together, and then make a shell script wrapper around the original binary, to set that env var
<Henson>
clever: thank you
<nervengift>
clever: do you know of a good example?
<clever>
nervengift: firefox and chrome both do that
<nervengift>
thank you! will have a look
stranger___ has quit [Quit: Connection closed for inactivity]
<{^_^}>
[nixpkgs] @worldofpeace pushed 2 commits to master: https://git.io/fj6wW
stepcut has joined #nixos
drakonis1 has joined #nixos
Okinan has joined #nixos
thc202 has quit [Ping timeout: 250 seconds]
ambro718 has quit [Quit: Konversation terminated!]
polman has quit [Ping timeout: 248 seconds]
ilmu has quit [Read error: No route to host]
polman has joined #nixos
phreedom_ has joined #nixos
phreedom has quit [Ping timeout: 260 seconds]
xcthulhu has quit [Quit: xcthulhu]
ericsagnes has quit [Ping timeout: 248 seconds]
linarcx has quit [Quit: WeeChat 2.5]
o1lo01ol1o has joined #nixos
o1lo01ol1o has quit [Ping timeout: 245 seconds]
jackdk has joined #nixos
xcthulhu has joined #nixos
ericsagnes has joined #nixos
<kraem>
Ralith: thanks - have been looking into home-manager but thought i'd learn "vanilla" nixos configurations first but cannot seem to solve the ssh-agent/gpg-agent/gnome-keyring situation so think i'll give it a try :)
<Ralith>
I don't think plain nixos tries very hard for user-level stuff
o1lo01ol1o has joined #nixos
alex`` has quit [Ping timeout: 248 seconds]
iqubic has joined #nixos
WilliButz has quit [Quit: WeeChat 2.5]
WilliButz has joined #nixos
Guest68314 has quit [Quit: WeeChat 2.4]
<iqubic>
Alright... Time to free up a ton of space by deleting all of Windows 10 from my system.
<iqubic>
How do I do that?
<slabity>
iqubic: Resize your partition?
<iqubic>
I haven't used Windows 10 in like a year.
<iqubic>
slabity: First I need to figure out which partitions Windows 10 is using, then I need to resize the current partitions.
<slabity>
Funny. I was just setting up a Windows 10 VM
<slabity>
You can use lsblk to find what each partition is.
<iqubic>
And I'm tearing down my Windows 10 + Nixos dual boot, because Windows 10 is taking up a butt load of space and don't like that.
wfranzini has quit [Remote host closed the connection]
<nouserpass>
I got around my issue by going to an older generation, so thanks NixOS ;)
<iqubic>
I don't have gparted installed.
arjen-jonathan has quit [Ping timeout: 252 seconds]
<pie_>
nouserpass, yay. id look into setting up a recovery boot as well if i was you. ( clever i still think we should have an option for enabling this :v
knupfer has quit [Ping timeout: 252 seconds]
<pie_>
havent had time to look at it tho
<iqubic>
And it turns out that I don't have enough space on my drive to install gparted either.
<iqubic>
This is why I'm looking to free up some space.
<{^_^}>
[nixpkgs] @aij opened pull request #64316 → nixos/zfs: enable requestEncryptionCredentials by default → https://git.io/fj6wd
<pie_>
iqubic, have you tried to do a garbage collection
<iqubic>
I have.
<iqubic>
I did that earlier today.
<clever>
iqubic: and what happens when you GC?
<slabity>
pie_: Would that just be setting up a new system profile and adding it to the bootloader?
<iqubic>
Well, I can't garbage collect any more.
<iqubic>
error: could not set permissions on '/nix/var/nix/profiles/per-user' to 1777: No space left on device
<pie_>
slabity, idk clever has some snippet i cannibalized from him :P
<pie_>
so, things to look at: clearning journald logs, garage collection (you did this), optimising the nix store, looking for garbage collection roots you dont care about anymore
<pie_>
actually do we not have a wiki page that covers all this already?
<clever>
iqubic: you have 32gig worth of snapshots for /home
<clever>
iqubic: what does `zfs list -t snapshot -r -o name,used,refer,written latitude-tank/home` output?
<iqubic>
1. remove the Windows Bootloader from my /boot partition. 2. Remove Windows entries from my boot menu. 3. Resize my ZFS tank so that it takes up all the available space.
<iqubic>
How do I do those things?
<clever>
sda6 is at the end of the disk, it cant be resized to be bigger
<iqubic>
Oh? So what should I do here?
<clever>
and sda5 is in the way of using the whole disk up properly
<iqubic>
I'm going to be deleting sda5.
<clever>
now or later?
<iqubic>
Now.
<iqubic>
So what can I do here to give more space to Linux?
<clever>
after its gone, you can create a new sda2, that takes up all of the space that you can
<clever>
that will create a mirror, and automatically copy all data from sda5 to sda2
<iqubic>
can't I just copy the file system from sda6 to sda2, and then delete sda6?
tmaekawa has joined #nixos
<clever>
thats what attach will do
<iqubic>
Ah Nice.
<clever>
watch `zpool status` to see when its done
<clever>
and then `zpool detach latitude-tank /dev/sda5` to un-mirror it, making sda2 the only part of the pool
<clever>
thats just step one, which will move the entire system from the end of the disk to the start
<iqubic>
How do I create a new partition with fdisk?
<clever>
`n`
<iqubic>
clever: step two is to go into a live environment and extend that partition, right?
<clever>
dont need a live environment
<iqubic>
I don't?
<clever>
you can resize the partition while its in-use i believe
<clever>
and the final step, is `zpool online -e latitude-tank /dev/sda2`
<clever>
which will expand the vdev within sda2 to occupy the new space you gave it
<iqubic>
I feel that I have to reboot because when I save the changes with fdisk, it gives me this message before closing down: http://dpaste.com/3MF4Y23
<clever>
yeah, if it gives an error like that, youll need to reboot
<clever>
but you dont need a livecd
<iqubic>
Alright sounds good.
<iqubic>
Before I reboot, I'd like to remove the windows bootloader from my esp partition.
<iqubic>
How do I do that?
<clever>
just find the windows specific files and delete them
<{^_^}>
[nixpkgs] @jb55 opened pull request #64317 → purple-instagram: init at 2019-01-13 → https://git.io/fj6rI
<iqubic>
And will that remove the entries from my boot menu too?
<clever>
for that, you need efibootmgr
<clever>
[root@system76:~]# efibootmgr -v
<iqubic>
Oh? Do I?
<clever>
iqubic: and check its man page
<iqubic>
Can
<iqubic>
Can't I just use os-prober?
<clever>
thats not what os-prober does
<iqubic>
Oh? What does os-prober do?
<clever>
os-prober detects other OS's and adds them to the grub config
<clever>
efibootmgr deals with efi vars that point to .efi files on the ESP
<iqubic>
Ah. I see.
<clever>
in theory (with a usable firmware), you can use efibootmgr entirely, and manage every OS in the firmware cfg
tmaekawa has quit [Quit: tmaekawa]
<clever>
but, when you move a disk to a new machine, the motherboard wont know about it, so its already a fail :P
<clever>
and you must boot via efi, to configure it to boot via efi!!
<clever>
and many firmware lack a menu, and only support a single os, lol
<clever>
some even ignore the efi vars, and only boot the windows efi file
<hexa->
removable firmware *shrug*
<clever>
"it boots windows, ship it"
<clever>
ive even heard of cases where you cant delete an efi var entry, if the ESP it mentions doesnt exist
<clever>
so if you delete the ESP (or loose the disk), your motherboard is basically bricked :P
<{^_^}>
[nixpkgs] @worldofpeace pushed 5 commits to master: https://git.io/fj6r1
leothrix has quit [Remote host closed the connection]
<gyroninja___>
Note that it timesout several minutes after the downloads have started
<gyroninja___>
so I assume I'm just taking too long to download these files
leothrix has joined #nixos
<clever>
gyroninja___: check with wireshark to see if its actualy sending data the whole time
<gyroninja___>
okay
alex`` has joined #nixos
<gyroninja___>
I'm seeing traffic with fastly
<gyroninja___>
and when it timed out it stopped
mbrgm_ has joined #nixos
shabius_ has joined #nixos
mbrgm has quit [Ping timeout: 252 seconds]
<clever>
gyroninja___: which end last sent a packet, and where any retries present?
mbrgm_ is now known as mbrgm
<iqubic>
That's a lot of space saved: "52730 store paths deleted, 36181.37 MiB freed"
<gyroninja___>
I'm not that well versed into how to use wireshark
<clever>
gyroninja___: you generally want to right click the packets for this stream, and hit follow tcp stream to filter it
<gyroninja___>
under statistics it right now shows Tx Bytes as 266M and Rx bytes as 6561k
shabius has quit [Ping timeout: 252 seconds]
<gyroninja___>
I would assume Rx would be what I am recieving, so that would be weird
<clever>
gyroninja___: you generally want to just look at the packets, you should see push packets from fastly, and ack packets going back for each one
<iqubic>
clever: You got your arguments backwards in the command you gave me.
<clever>
2019-07-04 19:10:36 < clever> watch `zpool status` to see when its done
mexisme_ has quit [Ping timeout: 245 seconds]
mexisme_ has joined #nixos
<iqubic>
Currently its resilvering.
<clever>
it will need to copy about 90gig of data
<iqubic>
How did you get that number?
<clever>
the size of your partitions
<iqubic>
Is that just the current size of the entire latitude-pool?
<clever>
/dev/sda6 294348800 500117503 205768704 98.1G Linux filesystem
<clever>
yes
<iqubic>
That's closer to 100G of data.
<gyroninja___>
clever: Okay, so in the final traffic before I get errors on the nix side, wireshark highlighting in red that I sent fastly a [RST, ACK] tcp packet and then it alternates between the server sending ACK and me sending RST which happens ~50 times
<iqubic>
Why is the process called resilvering?
v0|d has joined #nixos
<clever>
gyroninja___: fastly sending an rst means that the fastly server is trying to close the connection
<gyroninja___>
well I'm the source of the RST packets
<clever>
iqubic: its to do with fixing a physical mirror, replacing the silver on the back of it
<iqubic>
I just found this which states the same thing.
kraem has quit [Read error: Connection reset by peer]
<gyroninja___>
okay I'll try
zacts has quit [Quit: WeeChat 2.4]
MmeQuignon has quit [Ping timeout: 258 seconds]
<gyroninja___>
yeah wget worked
kraem has joined #nixos
<gyroninja___>
I ran it at the same time as nix trying to do it and nix failed to download it, but I succeeded
MmeQuignon has joined #nixos
<clever>
gyroninja___: how long did the download take?
<gyroninja___>
it might also be note worthy that nix is trying to download 2 packages at the same time
<gyroninja___>
which is why I tried the wget as the same time
<gyroninja___>
the download took 5 minutes and 16 seconds
<gyroninja___>
I'd say it took nix almost exactly 5 minutes to timeout
<gchristensen>
ipv6 or no?
<gyroninja___>
ipv4
<gyroninja___>
only
<gyroninja___>
Yep the "HTTP error 200 (curl error: Timeout was reached)" happens after exactly 5 minutes
<gyroninja___>
I timed it that time
phreedom_ has quit [Ping timeout: 260 seconds]
___laika has quit [Ping timeout: 258 seconds]
<iqubic>
Done resilvering.
noudle has quit []
<gyroninja___>
Invoking curl manually myself to download the file works fine and doesn't time out once I hit 5 minutes
utdemir has joined #nixos
hiroshi[m] has joined #nixos
hiroshi[m] has joined #nixos
hiroshi[m] has quit [Changing host]
<utdemir>
Hey. For some reason my Nix commands try to use "https://qfpl.cachix.org" as a substituer. I remember adding it and removing it once, however currently my configuration.nix does not reference it, and neither my /etc/nix.conf. Can anyone tell me where else it can come from?