Some steps to (hopefully) help you fix your Nix install on OS X after an upgrade.

a_help-osx-borked-my-nix.md

Apple Borked my Nix!

Ok, so you've had nix (home-manager) working fine. Then Apple tells you it's time to update.

Ok. Reboot. Oops. It has now broken your Nix setup. Here's some stuff to work through. YMMV.

Note: This is what worked for me, who was just using nix + home-manager. The upgrade that I last did that caused all these issues was 12.3.X > 12.4

Useful Links

• Install instructions for OS X - it has a good summary of what is being done during install. https://nixos.org/manual/nix/stable/installation/installing-binary.html#macos-installation-a-namesect-macos-installation-change-store-prefixaa-namesect-macos-installation-encrypted-volumeaa-namesect-macos-installation-symlinkaa-namesect-macos-installation-recommended-notesa
• Actual install script https://github.com/NixOS/nix/blob/master/scripts/install-darwin-multi-user.sh
• Actual install script that works with setting up the Nix volume. https://raw.githubusercontent.com/NixOS/nix/master/scripts/create-darwin-volume.sh

Nix setup will modify your /etc/zshrc. When OS X updates, it will sometimes restore the original /etc/zshrc.

Check the bottom of your etc/zshrc, you should see the following:

# Nix
if [ -e '/nix/var/nix/profiles/default/etc/profile.d/nix-daemon.sh' ]; then
  . '/nix/var/nix/profiles/default/etc/profile.d/nix-daemon.sh'
fi
# End Nix

If it is NOT there, you want to add this back to the bottom of your /etc/zshrc. Or you can try to future proof things a little, and add it to the top of your own ~/.zshrc.

Nix Store volume check

From terminal try diskutil list | grep Nix. You should see something like

7:                APFS Volume Nix Store               10.1 GB    disk3s7

So that means we have the volume. That last bit - disk3s7 is the disk identifier. So then try diskutil info disk3s7

   Device Identifier:         disk3s7
... snip ...
   Volume Name:               Nix Store
   Mounted:                   Yes
   Mount Point:               /nix
... snip ...
   Volume UUID:               53E764A4-140E-4980-9EE2-39CB2404DAFF
... snip ...
   FileVault:                 Yes
... snip ...

So if Mounted is No then you need to mount this. Probably means the auto-mount script is broken. Check if /Library/LaunchDaemons/org.nixos.darwin-store.plist exists. If not, you can recreate it with the provided script - generate-volume-mount-plist.sh. This script is based off the nix install script create-darwin-volume.sh.

USE THE VOLUME UUID you found above. And use unencrypted if your volume is FileVault: No.

NIX_VOLUME_UUID=53E764A4-140E-4980-9EE2-39CB2404DAFF NIX_VOLUME_STATE=encrypted ./generate-volume-mount-plist.sh

Check you have /nix

If your /nix root level is missing, that means the mount point of your Nix Store got hosed. Let's rebuild it.

• Check for /etc/synthetic.conf. If it exists, make sure you have an entry for nix. If you don't have one, then you will need to set it up. It should be as simple as a file that just says:

nix

• Check your /etc/fstab file. If it exists, check you have an entry for /nix. If it doesn't, you will need to create it. WITH YOUR UUID you found earlier.

#
# Warning - this file should only be modified with vifs(8)
#
# Failure to do so is unsupported and may be destructive.
#
UUID=53E764A4-140E-4980-9EE2-39CB2404DAFF /nix apfs rw,noauto,nobrowse,suid,owners

Done

Reboot.

generate-volume-mount-plist.sh

#!/usr/bin/env bash
set -eu
set -o pipefail

# Script to regenerate the automount for Nix Store to /nix
# Cobbled together looking at https://raw.githubusercontent.com/NixOS/nix/master/scripts/create-darwin-volume.sh
readonly NIX_ROOT="${NIX_ROOT:-/nix}"
readonly NIX_VOLUME_LABEL="${NIX_VOLUME_LABEL:-Nix Store}"
readonly NIX_VOLUME_MOUNTD_DEST="${NIX_VOLUME_MOUNTD_DEST:-/Library/LaunchDaemons/org.nixos.darwin-store.plist}"
readonly NIX_VOLUME_STATE=${NIX_VOLUME_STATE:-encrypted} # encrypted|unencrypted
readonly NIX_VOLUME_UUID="${NIX_VOLUME_UUID:-}"

ui_confirm() {

    local prompt="[y/n] "
    echo -n "$prompt"
    while read -r y; do
        if [ "$y" = "y" ]; then
            echo ""
            return 0
        elif [ "$y" = "n" ]; then
            echo "You said no. Exiting."
            return 1
        else
            echo "Sorry, I didn't understand. I can only understand answers of y or n"
            echo -n "$prompt"
        fi
    done
    echo ""
    return 1
}

uuid_set() {
    if [[ -z "$NIX_VOLUME_UUID" ]]; then
        echo "You need to set the value of NIX_VOLUME_UUID. You can find this info with 'diskutil info -all' or using Disk Utility app."
        return 1
    fi
    return 0
}


volume_state_set() {
    if [[ "$NIX_VOLUME_STATE" != "encrypted" && "$NIX_VOLUME_STATE" != "unencrypted" ]]; then
        echo "Invalid value for NIX_VOLUME_STATE - must be encrypted or unencrypted"
        return 1
    fi
    return 0
}


check() {
    uuid_set
    volume_state_set
    echo "All checks passed."
}

generate_mount_command() {
    local cmd_type="$1" # encrypted|unencrypted
    local volume_uuid mountpoint cmd=()
    printf -v volume_uuid "%q" "$2"
    printf -v mountpoint "%q" "$NIX_ROOT"

    case "$cmd_type" in
        encrypted)
            cmd=(/bin/sh -c "/usr/bin/security find-generic-password -s '$volume_uuid' -w | /usr/sbin/diskutil apfs unlockVolume '$volume_uuid' -mountpoint '$mountpoint' -stdinpassphrase");;
        unencrypted)
            cmd=(/usr/sbin/diskutil mount -mountPoint "$mountpoint" "$volume_uuid");;
        *)
            failure "Invalid first arg $cmd_type to generate_mount_command";;
    esac

    printf "    <string>%s</string>\n" "${cmd[@]}"
}

generate_mount_daemon() {
    local cmd_type="$1" # encrypted|unencrypted
    local volume_uuid="$2"
    cat <<EOF
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple Computer//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
  <key>RunAtLoad</key>
  <true/>
  <key>Label</key>
  <string>org.nixos.darwin-store</string>
  <key>ProgramArguments</key>
  <array>
$(generate_mount_command "$cmd_type" "$volume_uuid")
  </array>
</dict>
</plist>
EOF
}

test_voldaemon() {
    test -f "$NIX_VOLUME_MOUNTD_DEST"
}

task() {
    echo "$@"
}

_sudo() {
    shift # throw away the 'explanation'
    /usr/bin/sudo "$@"
}

setup_volume_daemon() {
    local cmd_type="$1" # encrypted|unencrypted
    local volume_uuid="$2"
    if ! test_voldaemon; then
        task "Configuring LaunchDaemon to mount '$NIX_VOLUME_LABEL'" >&2
        # Note: `-u NONE` disables vim plugins/rc; see note on --clean earlier
        _sudo "to install the Nix volume mounter" /usr/bin/ex -u NONE "$NIX_VOLUME_MOUNTD_DEST" <<EOF
:a
$(generate_mount_daemon "$cmd_type" "$volume_uuid")
.
:x
EOF

        # TODO: should probably alert the user if this is disabled?
        _sudo "to launch the Nix volume mounter" \
            launchctl bootstrap system "$NIX_VOLUME_MOUNTD_DEST" || true
        # TODO: confirm whether kickstart is necessesary?
        # I feel a little superstitous, but it can guard
        # against multiple problems (doesn't start, old
        # version still running for some reason...)
        _sudo "to launch the Nix volume mounter" \
            launchctl kickstart -k system/org.nixos.darwin-store
    fi
}


main() {
    check
    setup_volume_daemon $NIX_VOLUME_STATE $NIX_VOLUME_UUID
}

main

Nix never seemed to work properly on my machine, I didn't really understand why. After some digging online, I found this and I was able to fix my Nix installation. Thank you for compiling this list of things to check, it helped me out a lot!

Thanks v much for this! This just happened to me upgrading from 12.6.1 to 12.6.2 -- just the /etc/zshrc.

great! glad it helped

Thank you so much for this. Update to 12.6.4. Then fix /etc/zshrc everything went back.

Thanks you so much! I just upgraded macOS from 13.4 to 13.5 and my nix was gone. Followed this guide to restored it!