Lose one hour thinking this RAID controller driver must be buggy because « mount /dev/sda1 /mnt/sda1 » errors out with « No such device« . After this hour, figure out from man 2 mount that ENODEV means « filesystemtype not configured in the kernel. » instead. Duh.
Accept the fact that some gigabit switches from an anonymous vendor that has a name remarkably similar to Hell, switches that cost more than a thousand euros, just lose all their configuration after a power loss. Just reconfigure switches.
When replacing a failed RAID disk, either make the new disk’s partition table with »sfdisk -d /dev/sda | sfdisk /dev/sdb » (if you’re replacing sdb), or if you have to do it by hand, don’t forget to set the partitions type to fd (Linux RAID autodetection) instead of 83 (Linux). If you don’t, your partition will be kicked out of the array at each reboot.