<clever>
iqubic: did you use sdb1 or sdb previously?
<clever>
iqubic: what dd command did you previously use?
<clever>
iqubic: you can do that with just dd, and it must not be mounted when doing so
<clever>
iqubic: why are you using k3b on a usb stick?
<clever>
iqubic: mkfs
<clever>
iqubic: that filesystem doesnt support write
<clever>
iqubic: TYPE="iso9660"
<clever>
iqubic: 404
<clever>
iqubic: can you paste it here?
<clever>
iqubic: what does blkid say about the device your mounting?
<clever>
the output is supposed to be based entirely on the build instructions, and not when you built it
<clever>
not pure
<clever>
Akii: nix doesnt allow any timestamps in the store,all files are last-modified at unix epoch + 1
<clever>
d1rewolf: you want to see what the child proccesses of nixos-rebuild and nix-daemon are with `ps -eH x` and find one that is blocked on something
<clever>
thats one reason to use zfs with automatic snapshots
<clever>
so its always lost on reboot
<clever>
the whole /run is a tmpfs
<clever>
it will repair itself on every boot
<clever>
if you cant use ssh to get root, then your only option is to reboot
<clever>
tilpner: only the sudo in /run/wrappers/bin/sudo works
<clever>
tilpner: that sudo will never work
<clever>
tilpner: what binary did you run?
<clever>
i dont remember how i figured out those args, i just mashed random things together over a decade ago and memorized what worked, lol
<clever>
a literal x
<clever>
an argument to ps
<clever>
d1rewolf_: its no longer building anything at that point, check `ps -eH x`
<clever>
maurer: there is also the problem that having permission to create iam roles, basically gives you complete admin
<clever>
sophiag: can you pastebin all of the errors?
<clever>
sophiag: stateVersion cant corrupt the nix store
<clever>
sophiag: do you know what caused the corruption?
<clever>
iqubic: this is also a good use for my rescue env
<clever>
its meant to be an installer, not a livecd where you can demo things
<clever>
that would be simpler and safer
<clever>
ah
<clever>
and why do you need to shrink it while its running?
<clever>
try it first
<clever>
try it first
<clever>
iqubic: you may also want to try gparted, it supports resize and has a gui
<clever>
iqubic: fdisk -l
<clever>
(Modern Linux 2.6 kernels will support on-line resize for file systems mounted using ext3 and ext4; ext3 file systems will require the use of file systems with the resize_inode feature enabled.)
<clever>
iqubic: check its man page and read things closely
<clever>
it should detect that its mounted, and contact the kernel and arrange for things to happen
<clever>
iqubic: first, run resize2fs on the device, with the new size you want to use
<clever>
iqubic: depends on the filesystem and which direction the size is going
<clever>
abueide: if using efi, you have to set the removable flag when you install
<clever>
RetardedOnion: its plausible that its using cross-site scripting to connect to something like the freenode webchat from your ip, and then spam more
<clever>
i found this being said in another room
<clever>
2018-07-25 17:52:23 < swedneck[m]> it seems to be a virus that relies on shitty web clients to propagate
<clever>
rescue_boot.nix just saves you from needing to find that stick for most problems
<clever>
mikky: yeah, you would still need a usb stick then
<clever>
chances
<clever>
mikky: but if /boot is damaged, what are the changes the rootfs is ok?
<clever>
well, if /boot is damaged, itsless of a repair, and more of a reinstall
<clever>
you cant loose the /boot
<clever>
logzet: you can loose the usb stick :P
<clever>
drakonis: and /boot could become part of / if you wanted
<clever>
drakonis: using boot.loader.efi.efisysmountpoint you can have seperate partitions for /boot and /boot/efi
<clever>
it eats about 300-400mb in /boot though
<clever>
so as long as /boot is intact, you can repair the system, or even do a full reinstall
<clever>
logzet: this nixos module will auto-generate an entire nixos installer, and put it in /boot, along with a grub entry to select it
<clever>
logzet: i also make my /boot at least 1gig
<clever>
logzet: i think it uses something like ${out}-root as the rootfs for the build, with many bind mounts
<clever>
logzet: if sandboxing is on, it builds it in a chroot like namespace, if sandboxing is off, it just builds directly into /nix/store/
<clever>
logzet: it would let you have a small fat32 efisys, then a much larger (or even part of the rootfs) /boot for many kernels, and proper journaling
<clever>
so / can have automatic snapshots, but /nix doesnt
<clever>
logzet: i always use zfs, with /nix on its own dataset
<clever>
logzet: not really, as long as configuration.nix knows to mount it
<clever>
logzet: id just create a new /boot and /, then clone everything over while booted into a livecd
<clever>
logzet: depends on what the existing layout is, and what your filesystem preferences are
<clever>
gchristensen: ah, i think its +d and freenode doesnt have it
<clever>
gchristensen: i think there is a channel mode that forces you to idle for 30 seconds before you can speak
<clever>
too slow!
<clever>
logzet: and if the config is broken after copying the partitions, you could just use nixos-enter to chroot into it, fix the config, and nixos-rebuild boot
<clever>
logzet: but since its nix, you can also just copy /etc/nixos/configuration.nix and nixos-install again, and it will be virtually identical
<clever>
logzet: if you copy the partitions over while it is not running, it should be fine, as long as the old config in /boot is able to find the right partitions to mount
<clever>
keith_analog: if you do `buildInputs = [ gperftools ];` then the stdenv will add a -L flag for you
<clever>
fresheyeball: then re-run nix-build again
<clever>
fresheyeball: delete result, then run nix-store --delete on the output path it made
<clever>
fresheyeball: if its an intermitent failure due to ram, then you need to fight succeed on failure
<clever>
fresheyeball: then you need to change something to correct the failure, and re-run it
<clever>
fresheyeball: you need to check for result/nix-support/failed
<clever>
fresheyeball: ah, succeed on failure
<clever>
gchristensen: the is getting +o every time somebody joins the room
<clever>
gchristensen: does it need to be op'd every time somebody joins?
<clever>
fresheyeball: with `ls -l result result/`
<clever>
fresheyeball: can you post a copy from the terminal of you running it, ls'ing result, running it again, and ls'ing result again?
<clever>
fresheyeball: if you add a `ls -ltrh $out` after that line, does it show files being added to out?
<clever>
fresheyeball: what part of that is writing to $out?
<clever>
so it never runs the checkPhase
<clever>
fresheyeball: you dont have a doCheck = true;