Comment 3 for bug 1443542

i'm looking at http://bazaar.launchpad.net/~curtin-dev/curtin/trunk/view/head:/helpers/common#L330 (which admittedly i just committed to, but did not change this general behavior).

and i'm not really sure how it can get there...
partition_main() line 374 calls wipedev. wipedev wipes start and end, and then calls blockdev --rereadpt and udevadm settle.

then we partition via one of the pt_* things. and each of those do a 'blockdev --rereadpt "$target"' and 'udevadm settle.

that 'partition' is called from
 http://bazaar.launchpad.net/~curtin-dev/curtin/trunk/view/head:/curtin/commands/block_meta.py#L204
and then after that block_meta does the mkfs that apparently finds no /dev/sda1.