I'm setting up a new Ubuntu home server and ran into two issues. I created an 8 TB RAID 1 array with mdadm, but after reboot it doesn't appear to mount. Mounting it manually fails with "special device /dev/md0 does not exist," while the server reports "Can't lookup blockdev." I added entries to mdadm.conf and /etc/fstab and ran update-initramfs as suggested by the guide I followed. The guide's mdadm step used `mdadm --detail --scan` rather than manually entering the array information. The output of `cat /proc/mdstat` only shows the RAID personalities line and `unused devices: `, with no active array. Separately, the Pi-hole installer appears to hang before displaying its setup interface. What should I check to get the RAID array assembling and mounting, and what could cause the Pi-hole installation to stall?
3 Answers
The `/proc/mdstat` output indicates that no mdadm array is currently assembled, so `/dev/md0` hasn’t been created yet. Check whether the member disks are visible with `lsblk` or `fdisk -l`, then inspect their RAID metadata with `sudo mdadm --examine /dev/sdX` for each disk. If the metadata is intact, try assembling the array with `sudo mdadm --assemble --scan` and check `/proc/mdstat` again. Once `/dev/md0` exists, identify its filesystem with `lsblk -f` or `blkid`, create the mount point, and mount the filesystem. The fstab entry should generally use the filesystem UUID, not just the md device name. Don’t run `mdadm --create` unless you have confirmed the array metadata is gone and you have backups, since recreating it can overwrite data.
The mdadm configuration is probably not the first problem here. `mdadm --detail --scan` only reports arrays that mdadm can already detect; it does not create or repair an array. Compare the UUID reported by `mdadm --examine` with the ARRAY line in `/etc/mdadm/mdadm.conf`, check the spelling and syntax, then rebuild the initramfs with `sudo update-initramfs -u`. Also make sure the disks are connected consistently and that the array was actually created successfully before the reboot. If `mdadm --assemble --scan` reports that no arrays were found, the examine output and `lsblk` layout are needed to determine whether the wrong devices were used or the RAID metadata is missing.
The Pi-hole problem is separate from the RAID issue. Run the installer from a real terminal with a stable network connection, preferably over Ethernet, and watch for an error instead of assuming it is still working. Check that DNS and outbound HTTPS work, that the system is up to date, and that no existing service is already occupying ports 53, 80, or 443. You can also run the installer with shell tracing or download the script first so you can inspect where it stops. Avoid running multiple installer attempts until you know whether the first process is still active.
I had only checked `/proc/mdstat`, and it showed no active arrays. I’ll verify that both drives are detected and inspect their mdadm metadata before trying to assemble anything.