# Storage ## btrfs Btrfs is with ZFS the only filesystem that can detect and prevent silent-data-corruption when paired with ECC-RAM. Additionally you can create software-RAIDs very easily without any controller. We don’t use RAID 5 since restoring data is a pain. Also the prices for huge harddisks are now very low. Btrfs will only work satisfyingly with Kernel 4.4 and above. Thats why we are using Debian stretch/sid. The command: `btrfs fi show` will show all set up btrfs-filesystems and the free space on them. Every fs will handled with an UUID. `btrfs scrub start /mountpoint` Will scrub the targeted fs. This means the fs will check all block-checksums and correct errors automatically. When started once it will run until finished. Scrubbing will take a lot of time (100 mb/s). You should scrub the fs once every month. The status of the recent scrub can be checked with: `btrfs scrub status /mountpoint`. To create a complete new RAID or add a harddisk to an existing one you first have to format the drive with a GPT: `parted -s /dev/sdz mklabel gpt mkpart primary 1% 100%`. This command will use parted to create a GPT on sdz (you can read out the drive-letters with `lsblk`). You should restart the server after this step, because the new disk will likely not be detected correctly. After the restart you add the device to an existing RAID with: `btrfs device add /dev/sdz1 /` Here sdz1 getting added to the RAID `/`. You can create a complete new RAID 1 from multiple devices with: `mkfs.btrfs -d raid1 -m raid1 /dev/sdc1 /dev/sdd1 /dev/sde1 /dev/sdf1` After adding devices to RAIDs you always have to balance the data between them: `btrfs balance start -mconvert=raid1 -dconvert=raid1 /` ## Mounting the SAN storage The SAN - storage area network - is a network storage attached over 8 Gbit fibre channel. Currently we have 15 TB of storage mounted on harriet at `/SAN`. Follow \[this manual] to mount the external storage. Additional to the manual see [this error](http://pissedoffadmins.com/os/mount-unknown-filesystem-type-lvm2_member.html/) which will occur. To solve it you need to run: Scan logical volumes: ``` lvscan ``` Load device module: ``` modprobe dm-mod ``` Change the volumes that exist to active: ``` vgchange -ay ``` Mount the logical partition ``` mkdir -p /mnt/VolGroup00/LogVol00 mount /dev/VolGroup00/LogVol00 /mnt/VolGroup00/LogVol00 ``` \[this manual]: