One of my colleagues managed to accidentally run something like rm -rf /var/tmp/ * on a Solaris machine that was the mail server for the entire organisation.
After the command finished they realised that the inadvertent space in front of the asterisk meant that the command did slightly more damage than intended.
They were told to leave the machine running to be able to fix it from a backup, but they rebooted instead.
An open file is still usable even after it’s been deleted, so the kernel and shell were still up and running … before the reboot …
If I recall, it took weeks to fix, involving floppy disks, Sun engineers and much egg on face.
One of my colleagues managed to accidentally run something like
rm -rf /var/tmp/ *on a Solaris machine that was the mail server for the entire organisation.After the command finished they realised that the inadvertent space in front of the asterisk meant that the command did slightly more damage than intended.
They were told to leave the machine running to be able to fix it from a backup, but they rebooted instead.
An open file is still usable even after it’s been deleted, so the kernel and shell were still up and running … before the reboot …
If I recall, it took weeks to fix, involving floppy disks, Sun engineers and much egg on face.
The rm command REALLY needs to have the -f flag changed edit: to mean “ignore write protection” not “go turbo fuckery mode”
Yeah, except for root that is exactly how it works.
As root you are permitted to shoot your own foot and are expected to know how to aim.
Actually its the same as it overrides confirmation settings…