If you do pull the drives, make sure to pull them with the machine off, and make sure to insert them with the machine off. Not having the machine off would cause mdraid to fail the raid as the drives are removed, and not having it off when inserting them would make it activate them one at a time which may or may not quite work. I have done my updates without pulling the drives (at least 1-2 fedora reinstalls) and just made sure to not let anaconda mess with the larger raid disks. Without the conf file it will still find the raids, it just won't have the same md* names. I have also installed a disk elsewhere and made sure the dracut config was setup to include all drivers (change to hostonly=no). 6 does hostonly=no, 7 does hostonly=yes, not sure what 8 does by default. Good luck. On Tue, Dec 17, 2019 at 12:44 AM Benjammin2068 <benjammin2068@xxxxxxxxx> wrote: > > Hey everyone, > > I have a quick and (hopefully) easy question... > > > At some point, I'm going to update my CentOS 6.10 system to CentOS 8.x > > My biggest concern is dealing with the mdadm based RAID6 that holds my important data. > > SO.... What's the safest thing I could do to protect it when I do the upgrade (which is essentially a fresh install of v8) > > I'm assuming I should: > > * copy /etc/mdadm.conf someplace > * copy /etc/fstab someplace (for reference more than anything) > * PULL THE DRIVES FROM THE SYSTEM. > > * Once the new OS is installed, plug the drives back in. > * Patch mdadm.conf and fstab as needed ..... and..... > > Anything else? > > Thanks a bunch, > > -Ben >
