Comment 3 for bug 14187

A little wild guess here: maybe a "sleep 5" after the "kill -9" as it was
originally proposed in latest patch for cupsys.postinst (Comment #6 of Bug 11037)
would do the trick; or maybe an additional sleep somewhere else in that script?
 Maybe cupsys doesn't have the time to properly stop before its restart in some
configurations? (I know that I haven't experienced this problem on my computer
during the last cupsys upgrade)