When vmbuilder fails it leaks 2 loop devices. One I can use umount -l and then losetup -d to remove but the other is unremovable.
When I inspect the loop devices I get the output
/dev/loop0: [0900]:278134825 (/tmp/tmpjdqO5Q)
Where /tmp/tmpjdqO5Q is the image file created by vmbuilder.
When I inspect it with fuser -mu I get the output
dev/loop0: 1(root) 9784c(root) 9785c(root) 9786c(root) 9787c(root) 9788c(root) 9789c(root) 9798(root) 9799(root) 9801(daemon) 9802(root) 9803(root) 9805(root) 9811(root) 9837(root) 9838(root) 9841(root) 9857(trevor) 10682(oneadmin) 10683(oneadmin) 10702(oneadmin) 10725(oneadmin) 10735(oneadmin) 10749(oneadmin) 11220(root) 26638(root) 26643(root) 26659(trevor) 27272(root) 27273(root)
And looking at some of these processes (ps -e | grep 97) gives
9784 tty1 00:00:00 getty
9785 tty4 00:00:00 getty
9786 tty5 00:00:00 getty
9787 tty3 00:00:00 getty
9788 tty6 00:00:00 getty
9789 tty2 00:00:00 getty
9798 ? 00:00:00 cron
9799 ? 00:00:00 sshd
9801 ? 00:00:00 atd
9802 ? 00:00:00 upstart-udev-br
9803 ? 00:00:00 udevd
9805 ? 00:00:00 upstart-socket-
9811 ? 00:00:00 libvirtd
9837 ? 00:00:00 udevd
9838 ? 00:00:00 udevd
9840 ? 00:00:00 kworker/u:0
9841 ? 00:00:00 sshd
9857 ? 00:00:00 sshd
9858 pts/0 00:00:00 zsh
Is the parenthetical output of losetup a red herring (/tmp/tmpjdqO5Q)? Am I safe to delete the file /tmp/tmpjdqO5Q and not touch the loop device?
