On 5/1/2017 2:35 PM, Brian C. Lane wrote: > The only udev synchronization we have is for device-mapper devices.
Ahh right.. perhaps we need to remedy that? If our test suite has a
problem with parted exiting before the device shows up, I would think
that pretty much all callers of it would.
It looks like Debian has been carrying a patch for some time that runs
system("udevadm settle") in linux_disk_commit. It seems like there
should be a better way of doing that.
