/dev/null: Permission denied
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| | udev (Ubuntu) |
High
|
Scott James Remnant (Canonical) | ||
Bug Description
Binary package hint: udev
When I ssh to my server I see "permission denied messages" in my console:
[...]
-bash: /dev/null: Permission denied
-bash: /dev/null: Permission denied
-bash: /dev/null: Permission denied
-bash: /dev/null: Permission denied
-bash: /dev/null: Permission denied
-bash: /dev/null: Permission denied
[...]
I [Ctrl] + [c] out of this loop and get dropped to bash. Checking then the permissions on "/dev/null", I see "crw-------" owned by root:root. After a "# /etc/init.d/udev restart" the permissions are "crw-rw-rw-". I recently upgraded from stock Dapper to Edgy, and currently running "default" Edgy (2.6.17-10-386).
udev is configured to properly set the permissions on the file (the following is in '/etc/udev/
I still experience this problem on every reboot.
Attached is my "/var/log/udev".
"# udevtest /class/mem/null" gives:
main: looking at device '/class/mem/null' from subsystem 'mem'
udev_rules_
udev_db_get_device: no db file to read /dev/.udev/
udev_node_add: creating device node '/dev/null', major = '1', minor = '3', mode = '0666', uid = '0', gid = '0'
main: run: 'socket:
main: run: 'socket:
I don't know if this is related, but doing "# ls -al /dev/ | grep crw-------" gives
crw------- 1 root root 5, 1 2006-10-31 07:39 console
crw------- 1 root root 109, 0 2006-10-30 23:46 lvm
crw------- 1 root root 1, 3 2006-10-27 22:52 null
crw------- 1 root root 108, 0 2006-10-27 22:52 ppp
crw------- 1 root root 4, 1 2006-10-30 23:46 tty1
crw------- 1 root root 4, 2 2006-10-30 23:46 tty2
crw------- 1 root root 4, 3 2006-10-30 23:46 tty3
crw------- 1 root root 4, 4 2006-10-30 23:46 tty4
crw------- 1 root root 4, 5 2006-10-30 23:46 tty5
crw------- 1 root root 4, 6 2006-10-30 23:46 tty6
And "# ls-al /lib/udev/devices/ |grep crw-------" gives
crw------- 1 root root 5, 1 2006-10-27 22:52 console
crw------- 1 root root 1, 3 2006-10-27 22:52 null
crw------- 1 root root 108, 0 2006-10-27 22:52 ppp
| description: | updated |
Are you able to access the console of this server?
Could you provide the output of "ls /etc/rcS.d" and "ls /etc/rc2.d" for me?
Thanks
| Changed in udev: | |
| assignee: | nobody → keybuk |
| importance: | Undecided → Medium |
| status: | Unconfirmed → Needs Info |
| importance: | Medium → High |
| Ordou (khelle) wrote : | #3 |
I'm still able to access the console.
"# ls /etc/rcS.d":
README S27evms
S01mountkernfs.sh S30checkfs.sh
S01readahead S35mountall.sh
S02hostname.sh S36mountall-
S05keymap.sh S39readahead-
S06keyboard-setup S40networking
S07linux-
S08loopback S45waitnfs.sh
S10udev S46mountnfs-
S11mountdevsubfs.sh S49console-setup
S13pcmciautils S50hwclock.sh
S15module-
S17procps.sh S55pppd-dns
S20checkroot.sh S70screen
S22mtab.sh S70x11-common
S25brltty S80bootmisc.sh
S25mdadm-raid S85urandom
S26lvm S90console-
"# ls /etc/rc2.d":
K81cupsys S20makedev S24mythtv-backend S99mdadm-raid
README S20mysql S25mdadm S99module-
S05vbesave S20postfix S89anacron S99networking
S10powernowd.early S20rsync S89atd S99rc.local
S10sysklogd S20samba S89cron S99rmnologin
S11klogd S20slimserver S90binfmt-support S99stop-readahead
S13gdm S20smartmontools S91apache2 S99udev
S19mysql-ndb-mgm S20ssh S98usplash S99webmin
S20backuppc S20xinetd S99evms S99x11-common
S20dbus S21mysql-ndb S99loopback
S20hddtemp S23ntp-server S99lvm
Delete /etc/rc2.d/S99udev
| Changed in udev: | |
| status: | Needs Info → Rejected |
| Ordou (khelle) wrote : | #5 |
Thanks. This solved the problem.
| turncoat (uskokovic) wrote : | #6 |
I have the same problem, described in bug 53040, but I don't even have /etc/rc2.d/S99udev.
I only have /etc/rcS.d/S10udev. Should it be there?
I had /etc/rc2.d/S50udev.
I deleted it and now it works.
(Feisty)
| Dale Amon (amon) wrote : | #8 |
Same problem in karmic. There are no rc files for udev. I can delete and do the mknod to correct it but it goes away on reboot. I have gone through my root .bash* files and removed umask just in case. rm'ing the file, doing the mknod with 0666 and manually restarting through init.d scripts gives me 0644; doing the same and using service udev restart leaves it as 0666. However a reboot sets it back to 0660.
Ideas?
| jhthayer (jhthayer) wrote : | #9 |
Similar problem in karmic. In addition to the bash-null-


Attachment: /var/log/udev