Comment 22 for bug 1864787

This comment was ~suggested by Walter/wxl, it's intention is to go upstream.

All testing here is on existing system
dell [optiplex] 755 (c2d-e6850, 5gb, amd/ati radeon rv516/x1300/x1550)
using sda9 as /, and an sda8 as /home

sda8 is never formatted, just re-used. sda9 is formatted
logs created as done can be viewed at http://iso.qa.ubuntu.com/qatracker/milestones/408/builds/210806/testcases/1701/results/

Following include a reboot between each test

Failures:
re-use sda8, format sda9 reiserf (comment #20 log)
re-use sda8, format sda9 ext4 (comment #21 verbose log)

re-use sda8, format sda9 ext4 using updated packages (log)

re-use sda8, format sda9 ext4 using updated packages-2 (log)
re-use sda8, format sda9 ext4 using updated packages-2 (verbose log)

re-use sda8, format sda9 reiserfs
re-use sda8, format sda9 xfs
re-use sda8, format sda9 btrfs

On the following failures, sda9 was deleted using KDE partition manager so sda9 didn't exist except as unallocated space before test

re-use sda8, format xfs
re-use sda8, format btrfs

refer to http://iso.qa.ubuntu.com/qatracker/milestones/408/builds/210806/testcases/1701/results/ for real results though & more details.

To get it working, I reboot, use KDE Partition Manager to delete the 'unknown' partition that exists after all aforementioned FAILED tests, and create it in whichever format I want (ext4, btrfs were tested today & worked), deactivate swap & exit KDE Partition Manager. Calamares can then use the prepared sa9; and it'll install perfectly.