Support #447
Updated by Daniel Curtis over 10 years ago
Like many developers, I like Linux; particularly Arch Linux. And like many sysadmins I like BSD, particularly FreeBSD. This is a guide of how I setup my recent developer laptop, it consists of a few goodies like: # ZFS # BlackArch PenTesting Distro # LUKS Emergency Self-Destruct # USB Boot Loader Additional enhancements may come later. However getting all of this goodness onto a computer takes a little bit of patience and understanding. h2. Securely Wipe the Hard Drive * Once booted into an Arch Live ISO, run the following to find the drive to erase: <pre> fdisk -l </pre> #* Now erase the primary hard drive, this guide uses /dev/sda for the primary hard drive: <pre> dd if=/dev/zero of=/dev/sda bs=4M </pre> #* Now erase the USB drive, this guide uses /dev/sdc for the USB Boot Loader drive: <pre> dd if=/dev/zero of=/dev/sdc bs=1M </pre> h2. Adding the repository and install ZFS The maintainer of ZFS on Arch has a signed repository that you can add to the @/etc/pacman.conf@. * Add the *[demz-repo-archiso]* repo: <pre> vi /etc/pacman.conf </pre> #* And run the following to add the [demz-repo-archiso] to the end of /etc/pacman.conf: <pre> echo "[demz-repo-archiso]" >> /etc/pacman.conf echo "Server = http://demizerone.com/$repo/$arch" >> /etc/pacman.conf </pre> * Now repo key needs to be received and locally signed: <pre> pacman-key -r 0EE7A126 pacman-key --lsign-key 0EE7A126 </pre> * Now update the repository information: <pre> pacman -Sy </pre> * Its time to install ZFS: <pre> pacman -S zfs </pre> * Load the ZFS kernel module: <pre> modprobe zfs </pre> * Check to see that the module was loaded: <pre> lsmod | grep zfs </pre> h2. Install the patched cryptsetup * Install the base-devel and libutil-linux packages: <pre> pacman -S base-devel libutil-linux </pre> * Grab the patch cryptsetup from the AUR <pre> mkdir ~/src && cd ~/src wget https://aur.archlinux.org/packages/cr/cryptsetup-nuke-keys/cryptsetup-nuke-keys.tar.gz tar xzf cryptsetup-nuke-keys.tar.gz cd cryptsetup-nuke-keys </pre> * Install cryptsetup <pre> makepkg -s PKGBUILD makepkg -i PKGBUILD y y </pre> h2. Preparing the USB Boot Loader * Find where the USB drive is by running: <pre> fdisk -l </pre> *NOTE*: Since I am using an Arch ISO from a USB drive, this guide will use /dev/sdc for the USB Boot Loader. * Open cfdisk on the UDB drive: <pre> cfdisk /dev/sdc </pre> #* Erase all partitions, create a small partition for the bootloader, then add a partition for the rest of the drive, for storage: <pre> [New] primary 512 [Bootable] (make sure to have sda1 selected) (Select Free Space) [New] primary (Rest of the USB space) [Write] yes [Quit] </pre> * Make an ext3 partition for @/boot@: <pre> mkfs.ext3 /dev/sdc1 </pre> * Make a FAT partition for general storage on the USB drive: <pre> mkfs.fat /dev/sdc2 </pre> h2. Setting up the encrypted hard drive * Create a LUKS volume on /dev/sda <pre> cryptsetup -i 15000 -c aes-xts-plain:sha512 -y -s 512 luksFormat /dev/sda </pre> > Enter passphrase: > Verify passphrase: * Add the LUKS Emergency Self-Destruct passphrase: <pre> cryptsetup luksAddNuke /dev/sda </pre> > Enter any existing passphrase: (existing password) > Enter new passphrase for key slot: (set the nuke password) > Verify passphrase: (verify the nuke password) * Open the LUKS volume: <pre> cryptsetup luksOpen /dev/sda root </pre> *NOTE*: This will create the mapped device */dev/mapper/root*. This is where the ZFS root will be installed. *# (Optional) Create a backup of the LUKS Header <pre> luksHeaderBackup /dev/sda --header-backup-file /path/to/backup-luksHeader.img </pre> *# (Optional) Restore the LUKS Header from a backup luksHeaderRestore /dev/sda --header-backup-file /path/to/backup-luksHeader.img h3. Preparing the encrypted primary hard drive * Open cfdisk on the primary hard drive: <pre> cfdisk /dev/mapper/root </pre> #* Add the primary partition for ZFS <pre> (Select Free Space) [New] primary (All of the HD space) [Type] BF [Write] yes [Quit] </pre> h2. Setting up the ZFS filesystem * Create the zpool: <pre> zpool create zroot /dev/mapper/root </pre> *WARNING*: Always use id names when working with ZFS, otherwise import errors will occur. * Create necessary sub-filesystem mount points such as /home and /vms can be created with the following commands: <pre> zfs create zroot/home -o mountpoint=/home zfs create zroot/vms -o mountpoint=/vms </pre> *NOTE*: That if you want to use other datasets for system directories (/var or /etc included) your system will not boot unless they are listed in /etc/fstab! We will address that at the appropriate time in this tutorial. h2. Swap partition ZFS does not allow the use swapfiles, but it is possible to use a ZFS volume as swap partition. It is important to set the ZVOL block size to match the system page size; for x86 and x86_64 systems that is 4k. * Create a 2 GB (or whatever is required) ZFS volume: <pre> zfs create -V 2G -b 4K zroot/swap </pre> * Initialize and enable the volume as a swap partition: <pre> mkswap /dev/zvol/zroot/swap swapon /dev/zvol/zroot/swap </pre> * Make sure to unmount all ZFS filesystems before rebooting the machine, otherwise any ZFS pools will refuse to be imported: <pre> zfs umount -a </pre> h2. Configure the ZFS root filesystem filesystemarch linux luks * First, set the mount point of the root filesystem: <pre> zfs set mountpoint=/ zroot </pre> *# and optionally, any sub-filesystems: <pre> zfs set mountpoint=/home zroot/home zfs set mountpoint=/vms zroot/vms </pre> * Set the bootfs property on the descendant root filesystem so the boot loader knows where to find the operating system. <pre> zpool set bootfs=zroot zroot </pre> * Turn off swap, if enabled: <pre> swapoff -a </pre> * Export the pool: <pre> zpool export zroot </pre> *WARNING*: Do not skip this, otherwise you will be required to use -f when importing your pools. This unloads the imported pool. *NOTE*: This might fail if you added a swap partition above. Need to turn it off with the @swapoff@ command. * Finally, re-import the pool: <pre> zpool import -d /dev/mapper -R /mnt zroot </pre> *NOTE*: @-d@ is not the actual device id, but the @/dev/mapper@ directory containing the symbolic links. If there is an error in this step, you can export the pool to redo the command. The ZFS filesystem is now ready to use. * Be sure to bring the @zpool.cache@ file into your new system. This is required later for the ZFS daemon to start. <pre> mkdir -p /mnt/etc/zfs cp /etc/zfs/zpool.cache /mnt/etc/zfs/zpool.cache </pre> *# If you don't have /etc/zfs/zpool.cache, create it: <pre> zpool set cachefile=/etc/zfs/zpool.cache zroot </pre> h2. Installing Arch * Start by mounting the boot partition <pre> mkdir /mnt/boot mount /dev/sdc1 /mnt/boot </pre> * Now change the repository to *demz-repo-core* <pre> vi /etc/pacman.conf </pre> #* And change @[demz-repo-archiso]@ to the following > [demz-repo-core] > Server = http://demizerone.com/$repo/$arch * Then install the base system <pre> pacstrap -i /mnt base base-devel grub openssh zfs </pre> * Generate the fstab for filesystems, use: <pre> genfstab -U -p /mnt | grep boot >> /mnt/etc/fstab </pre> * Edit the @/etc/fstab@. If you chose to create datasets for system directories, keep them in this fstab! <pre> vi /mnt/etc/fstab </pre> #* +Comment out the lines+ for the /, /root, and /home mountpoints, rather than deleting them. You may need those UUIDs later if something goes wrong. Anyone who just stuck with the guide's directions can delete everything except for the swap file and the boot/EFI partition. It seems convention to replace the swap's uuid with /dev/zvol/zroot/swap. #* Edit @/mnt/etc/fstab@ to ensure the swap partition is mounted at boot: <pre> vi /mnt/etc/fstab </pre> > /dev/zvol/zroot/swap none swap defaults 0 0 * Setup the initial environment: <pre> arch-chroot /mnt </pre> #* Set a root password <pre> passwd </pre> #* Set a hostname <pre> echo "archzfs" > /etc/hostname </pre> #* Set a local time <pre> ln -s /usr/share/zoneinfo/America/Los_Angeles /etc/localtime </pre> #* Set a local language by uncommenting *en_US.UTF-8* in @/etc/locale.gen@, then running: <pre> locale-gen </pre> #* Set a wired network connection <pre> cp /etc/netctl/examples/ethernet-dhcp /etc/netctl/wired netctl enable wired </pre> #* Set SSH to start at boot <pre> systemctl enable sshd.service </pre> h3. LXDE * Install the LXDE desktop <pre> pacman -S lxde xorg xorg-xinit dbus gvfs gvfs-smb echo 'exec startlxde' >> ~/.xinitrc startx </pre> h3. Add an administive user * It is generally a good idea not to run command directly as root, but rather as an administrative user using the sudo wrapper command * First install sudo: <pre> pacman -S sudo </pre> * And create a user: <pre> useradd -m -g users -s /bin/bash bob </pre> * Add bob to the sudoers file: visudo bob ALL=(ALL) ALL h2. Setup the bootloader and initial ramdisk When creating the initial ramdisk, first edit @/etc/mkinitcpio.conf@ and add *zfs* +before+ *filesystems*. Also, move *keyboard* hook +before+ *zfs* so you can type in console if something goes wrong; and also put *usb* +before+ *keyboard* and *encrypt* +before+ *zfs*. You may also remove -fsck- (if you are not using Ext3 or Ext4). * The @HOOKS@ line should look something like this: > HOOKS="base udev autodetect modconf block *usb keyboard encrypt zfs* filesystems" * Regenerate the initramfs with the command: <pre> mkinitcpio -p linux </pre> h2. Install and configure GRUB * Install GRUB to the primary hard drive: <pre> grub-install --target=i386-pc --recheck --debug /dev/sdc </pre> h3. Edit GRUB to boot off of the zroot pool grub-mkconfig does not properly detect the ZFS filesystem, so it is necessary to edit grub.cfg manually. * Edit the GRUB config: <pre> /boot/grub/grub.cfg </pre> #* Add or modify it similar to the following <pre> set timeout=2 set default=0 # (0) Arch Linux menuentry "Arch Linux" { set root=(hd0,msdos1) linux /vmlinuz-linux cryptdevice=/dev/sda:root root=/dev/mapper/root zfs=zroot rw initrd /initramfs-linux.img } </pre> h3. Finish the setup process * Exit the chroot environment: <pre> exit </pre> *Unmount all ZFS mount points: zfs umount -a * Unmount the bootloader partition: <pre> umount /mnt/boot </pre> * Export the zpool: <pre> zpool export zroot </pre> * Reboot: <pre> reboot </pre> h2. After the first boot If everything went fine up to this point, your system will boot. Once. For your system to be able to reboot without issues, you need to enable the zfs.target to auto mount the pools and set the hostid. * For each pool you want automatically mounted execute: <pre> zpool set cachefile=/etc/zfs/zpool.cache <pool> </pre> * Enable the target with systemd: <pre> systemctl enable zfs.target </pre> When running ZFS on root, the machine's hostid will not be available at the time of mounting the root filesystem. There are two solutions to this. You can either place your spl hostid in the kernel parameters in your boot loader. For example, adding *spl.spl_hostid=0x00bab10c*, to get your number use the hostid command. * The other, and suggested, solution is to make sure that there is a hostid in /etc/hostid, and then regenerate the initramfs image. Which will copy the hostid into the initramfs image. To do write the hostid file safely you need to use a small C program: <pre> #include <stdio.h> #include <errno.h> #include <unistd.h> int main() { int res; res = sethostid(gethostid()); if (res != 0) { switch (errno) { case EACCES: fprintf(stderr, "Error! No permission to write the" " file used to store the host ID.\n" "Are you root?\n"); break; case EPERM: fprintf(stderr, "Error! The calling process's effective" " user or group ID is not the same as" " its corresponding real ID.\n"); break; default: fprintf(stderr, "Unknown error.\n"); } return 1; } return 0; } </pre> * Copy it, save it as @writehostid.c@ and compile it with: <pre> gcc -o writehostid writehostid.c </pre> #* Finally execute it and regenerate the initramfs image: <pre> ./writehostid mkinitcpio -p linux </pre> You can now delete the two files writehostid.c and writehostid. Your system should work and reboot properly now. h2. Installing BlackArch h3. Add the *[Multilib]* repository * Make sure to uncomment the *[Multilib]* repo, similar to the following: > [Multilib] > Include = /etc/pacman.d/mirrorlist * Refresh pacman: <pre> pacman -Syy </pre> h3. Setting up as an Unofficial User Repository BlackArch is compatible with normal Arch installations. It acts as an unofficial user repository. # Run the strap.sh script from http://blackarch.org/strap.sh as root: <pre> curl -s http://blackarch.org/strap.sh | sudo sh </pre> # Run the following to add the BlackArch reposiroty to @/etc/pacman.conf@: <pre> echo "[blackarch]" >> /etc/pacman.conf echo "Server = http://mirror.team-cymru.org/blackarch/\$repo/os/\$arch" >> /etc/pacman.conf </pre> # Now run: <pre> pacman -Syyu </pre> h2. Installing other developer tools and packages There are a few more packages that I use in my day-to-day tasks, for brevity I will refer over to Issue #410. h2. Optimizing and Tweaking ZFS offers many useful features like snapshotting, replication, and dataset customization. Virtual Machine Optimizations Since I will be running virtual machines from my developer laptop, I will need a ZFS dataset that will enable/disable certain features like compression to allow VMs to run more smoothly. This is why the @zroot/vms@ dataset was created during the initial setup * Options for zfs can be displayed using the zfs command: <pre> sudo zfs get all zroot/vms </pre> #* This will return: <pre> NAME PROPERTY VALUE SOURCE zroot/vms type filesystem - zroot/vms creation Sun Aug 31 14:47 2014 - zroot/vms used 30K - zroot/vms available 29.5G - zroot/vms referenced 30K - zroot/vms compressratio 1.00x - zroot/vms mounted yes - zroot/vms quota none default zroot/vms reservation none default zroot/vms recordsize 128K default zroot/vms mountpoint /vms local zroot/vms sharenfs off default zroot/vms checksum on default zroot/vms compression on default zroot/vms atime on default zroot/vms devices on default zroot/vms exec on default zroot/vms setuid on default zroot/vms readonly off default zroot/vms zoned off default zroot/vms snapdir hidden default zroot/vms aclinherit restricted default zroot/vms canmount on default zroot/vms xattr on default zroot/vms copies 1 default zroot/vms version 5 - zroot/vms utf8only off - zroot/vms normalization none - zroot/vms casesensitivity sensitive - zroot/vms vscan off default zroot/vms nbmand off default zroot/vms sharesmb off default zroot/vms refquota none default zroot/vms refreservation none default zroot/vms primarycache all default zroot/vms secondarycache all default zroot/vms usedbysnapshots 0 - zroot/vms usedbydataset 30K - zroot/vms usedbychildren 0 - zroot/vms usedbyrefreservation 0 - zroot/vms logbias latency default zroot/vms dedup off default zroot/vms mlslabel none default zroot/vms sync standard default zroot/vms refcompressratio 1.00x - zroot/vms written 30K - zroot/vms logicalused 15K - zroot/vms logicalreferenced 15K - zroot/vms snapdev hidden default zroot/vms acltype off default zroot/vms context none default zroot/vms fscontext none default zroot/vms defcontext none default zroot/vms rootcontext none default zroot/vms relatime off default </pre> * The above output shows that compression is turned on, to disable the compression feature, while still keeping it active for all other datasets in the zpool, run: <pre> zfs set compression=off zroot/vms </pre> h3. Snapshotting ZFS Snapshot Manager The zfs-snap-manager package from AUR provides a python service that takes daily snapshots from a configurable set of ZFS datasets and cleans them out in a "Grandfather-father-son" scheme. It can be configured to e.g. keep 7 daily, 5 weekly, 3 monthly and 2 yearly snapshots. * First install some dependencies: *# python2-daemon <pre> mkdir ~/src && cd ~/src wget https://aur.archlinux.org/packages/py/python2-daemon/python2-daemon.tar.gz tar xzf python2-daemon.tar.gz cd python2-daemon makepkg -s PKGBUILD && makepkg -i PKGBUILD </pre> *# mbuffer <pre> cd ~/src wget https://aur.archlinux.org/packages/mb/mbuffer/mbuffer.tar.gz tar xzf mbuffer.tar.gz cd mbuffer makepkg -s PKGBUILD && makepkg -i PKGBUILD </pre> * Install zfs-snap-manager from the AUR <pre> cd ~/src wget https://aur.archlinux.org/packages/zf/zfs-snap-manager/zfs-snap-manager.tar.gz tar xzf zfs-snap-manager.tar.gz cd zfs-snap-manager makepkg -s PKGBUILD && makepkg -i PKGBUILD </pre> * Create a simple snapshot config <pre> vi /etc/zfssnapshotmanager.cfg </pre> #* And enter in the following: <pre> [zroot] mountpoint = / time = 21:00 snapshot = True schema = 7d3w11m5y </pre> * Create a simple systemd service configuration vi /etc/systemd/system/zfs-snap-manager.service #* And enter in the following: <pre> [Unit] Description=ZFS Snapshot Manager After=syslog.target [Service] Type=simple User=root Group=root WorkingDirectory=/usr/lib/zfs-snap-manager/ PIDFile=/var/run/zfs-snap-manager.pid ExecStart=/usr/lib/zfs-snap-manager/manager.py start ExecStop=/usr/lib/zfs-snap-manager/manager.py stop [Install] WantedBy=multi-user.target </pre> * And finally enable the service at boot and start the service: systemctl enable zfs-snap-manager.service systemctl start zfs-snap-manager.service The package also supports configurable replication to other machines running ZFS by means of zfs send and zfs receive. If the destination machine runs this package as well, it could be configured to keep these replicated snapshots for a longer time. This allows a setup where a source machine has only a few daily snapshots locally stored, while on a remote storage server a much longer retention is available. h2. Resources * https://wiki.archlinux.org/index.php/ZFS * https://wiki.archlinux.org/index.php/Installing_Arch_Linux_on_ZFS * http://blackarch.org/download.html