Hi, Dan Purgert wrote: > You "may" need to wait for the drive to finish reading the disc metadata > (i.e. drive light stops flashing) before mount(1) will not complain > about the lack of media. Not 100% sure if it's a generic problem, or > just my crappy USB CD drive :)
It's a Linux kernel problem. I have seen it with a dozen different drives since Debian 8 of 2015, kernel 3.x. The kernel does not wait after emitting a START/STOP UNIT command until the drive reports success as answer to TEST UNIT READY commands. I diagnosed it in 2020 and submitted a patch. https://lore.kernel.org/linux-scsi/20201006094026.1730-1-scdbac...@gmx.net/T/#u But my social competence does not suffice for successful kernel work. (Also my research of kernel commits does not hold any more. 2bbea6e11735 from march 2008 cannot be found. A commit from march 2018 has the same title "cdrom: do not call check_disk_change() inside cdrom_open()" but a different commit id. It is too young for causing the problem back in Debian 8.) Jörg-Volker Peetz wrote: > In my experience, `cdrskin -eject` works more reliable than `eject`. At least it waits for the drive to be ready before it orders it to become unready. Note that cdrskin -eject without dev= argument will first grope all optical drives of the computer in order to get a list of them and then chose the first one, which is most probably /dev/sr0 on a Linux system. This may last some time even if there is only one such drive. Consider to give the desired drive device file explicitely: cdrskin dev=/dev/sr0 -eject Have a nice day :) Thomas
