More precisely: The new subnet mask is ignored on a dhcp
refresh.
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Package: xserver-xephyr
Version: 2:1.7.7-10
Severity: wishlist
Ubuntu provides a much more detailed and up-to-date
man page for Xephyr, e.g.
http://manpages.ubuntu.com/manpages/lucid/man1/xserver-xephyr.1.html
The Author section even claims it was written for the
Debian project.
Do you think
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Package: coreutils
Version: 8.5-1
According to the man page
echo 'a b' | tr -s 'z' ' '
should return
a b
instead of
a b
since there is no space character in set1, but just a 'z'.
Regards
Harri
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Package: base-files
Version: 6.1
Severity: wishlist
Would it be possible to put /var/run and /var/lock (and maybe
others) into /dev/shm? This could help to avoid a lot of write
operations on /var, which would be interesting for SSDs. It
could also
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Package: ifupdown
Version: 0.6.10
If /etc/network/run does not exist, then ifupdown fails at
boot time.
I am not sure that writing run time information to /etc is OK.
AFAICS there is no package providing /etc/network/run, so
ifupdown seems to be
Package: openssh-client
Version: 1:5.5p1-6
Severity: wishlist
It would be very nice if ssh -X could print an appropriate
warning if there is no /usr/bin/xauth at the other end.
Currently it is only shown in verbose mode.
Regards
Harri
--
To UNSUBSCRIBE, email to
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Package: mg
Version: 20110120-1
Accessing http://www.xs4all.nl/~hanb/software/mg/ (mentioned in debian/\
copyright) I just got a 404.
The new link seems to be
http://homepage.boetes.org/software/mg/
Regards
Harri
-BEGIN PGP
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Package: clamav-milter
Version: 0.97+dfsg-2
Severity: wishlist
Running dpkg-reconfigure clamav-milter I wonder if the config window
about /tmp could tell how much space it approximately needs?
I got /tmp in tmpfs, so temp space is restricted.
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Package: clamav-milter
Version: 0.97+dfsg-2
dpkg-reconfigure clamav-milter gave me a config file clamav-milter
could not read. Error message:
Replacing config file /etc/clamav/clamav-milter.conf with new version
Starting Sendmail milter plugin for
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Package: debianutils
Version: 3.4.3
Building a custom kernel package fails with
make[1]: Entering directory
`/var/tmp/build_kernel/raw-kernel-2.6.37/linux-2.6.37'
sh
/var/tmp/build_kernel/raw-kernel-2.6.37/linux-2.6.37/arch/x86/boot/install.sh
Package: vlc
Version: 0.9.6-1
I get an error message saying
[0001] main libvlc error: cannot open config file
(/home/harri/.config/vlc/vlcrc): Not a directory
According to the FAQs (http://www.videolan.org/doc/faq/en/index.html#id245329)
this is not supposed to happen, but actually I
Does it have support for dhcp authentication (RFC 3118)?
Regards
Harri
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Package: grub-pc
Version: 1.98~20100126-1
Since the most recent update the usual set gfxpayload=keep does not
work anymore. After the boot time message loading ramdisk there is
no further screen output on the console. The PC boots, though.
Moving back to the previous grub version seems to fix
For testing I have kicked out the notail for the root
partition, rebooted, installed the old version 20100101,
ran grub-install and update-grub, removed /boot/grub/grubenv,
ran grub-editenv create, and rebooted again:
No problem.
Regards
Harri
signature.asc
Description: OpenPGP digital
PS: For an easy recovery in this special case it seems to be
sufficient to boot from CD, mount the reiserfs filesystem
with notail, and recreate /boot/grub/grubenv, e.g.
mount -o notail /dev/sda1 /mnt
mv /mnt/boot/grub/grubenv /mnt/boot/grub/grubenv.bak
cp -p
Package: grub-pc
Version: 1.98~20100128-1
Severity: wishlist
Do you think it would be possible to release new versions
of grub2 on experimental? I understand that the unstable
version is what the name says, but even packages much less
important than grub2 make use of the experimental section
to
Package: grub-pc
Version: 1.98~20100128-1
Severity: grave
After the upgrade I get the message
GRUB loading.
Welcome to GRUB!
error: the symbol 'grub_env_find' not found.
Entering rescue mode...
on the next reboot. The boot menu is not shown.
Moving back to
Hi Felix,
On 01/30/10 11:01, Felix Zielcke wrote:
Please check with `echo GET grub-pc/install_devices |
debconf-communicate' what device is stored in there.
Stored in _where_?
Probable it's not the one where your BIOS boots from. Or the kernel/udev
decided to name it different when
On 01/30/10 11:08, Felix Zielcke wrote:
I'm not so sure I get the point of your report, because it looks more
like a general request then a specific request to update the version in
there.
Maybe you just missed that we already have packages in experimental?
See e.g.
Package: xserver-xorg-core
Version: 2:1.7.4-2
If I set DontZap to false, then Ctrl-Alt-Backspace still
does not zap. The Xwindow server does not terminate.
The whole xorg.conf is:
Section ServerFlags
Option DontZap false
EndSection
Attached you can find the
On 01/31/10 16:37, Brice Goglin wrote:
Do you have terminate(ctrl_alt_bksp) in the output of setxkbmap -print ?
Nope.
See /etc/default/keyboard or /etc/default/console-setup, you probably
missed the question during upgrade.
It was a fresh install of xorg plus dependencies. There was
On 02/01/10 00:21, Debian Bug Tracking System wrote:
This is an automatic notification regarding your Bug report
which was filed against the xserver-xorg-core package:
#567808: DontZap false does not work anymore
It has been closed by Julien Cristau jcris...@debian.org.
Sorry, but the
Colin Watson wrote:
UUIDs are properties of filesystems, not disks or partitions, and the
MBR of a hard disk does not have a UUID. For disks, the choices for
stable identifiers are based on either the hardware serial number or the
bus path taken to reach the disk.
What about block device
Package: libvirt0
Version: 0.7.2-3
After the migration the guest is defined on both old and
new host. Esp. when you are dealing with a large set of
hosts there is a high risk that you activate both guests
by accident, corrupting the virtual disk image.
An automatic undefine would be a big
Sorry, but I a not sure that I got you correctly. I just
select the guest in the virt-manager gui, click on migrate,
and select the new host. I do not know whats happening
inside, but I have to assume that it is using libvirt.
Do you think this is a problem of virt-manager?
Regards
Harri
Guido Günther wrote:
The guest should be powered off but still defined - that's the correct
behaviour. Undefining the vm is out of question since you'd lose the
configuration data then.
I don't see this. The data has been migrated to the new host.
What would we loose?
IMHO having 2 virtual
Package: sendmail
Version: 8.14.3-9
During the sendmail first-time installation I got
Errors in generating sendmail.cf
*** ERROR: FEATURE() should be before MAILER()
*** MAILER(`local') must appear after FEATURE(`always_add_domain')*** ERROR:
FEATURE() should be before MAILER()
***
Package: installation-reports
Boot method: netinst iso booted via virt-install
Image version:
http://cdimage.debian.org/cdimage/daily-builds/daily/arch-latest/amd64/iso-cd/debian-testing-amd64-netinst.iso
Date: 20091205
Machine: kvm
Processor: Qemu Virtual CPU version 0.11.0
Memory: 1024 MByte
Package: installation-reports
Boot method: netinst iso booted via virt-install
Image version:
http://cdimage.debian.org/cdimage/daily-builds/daily/arch-latest/amd64/iso-cd/debian-testing-amd64-netinst.iso
Date: 20091206
Machine: kvm
Processor: Qemu Virtual CPU version 0.11.0
Memory: 1024 MByte
PS: I missed to mention the virt-install command line. Here it is:
virt-install --connect qemu:///system -n squeeze_x -r 1024
--os-variant=debiansqueeze -v -c
/export/storage/isos/debian-testing-amd64-netinst.iso --disk
path=/export/storage/squeeze_x.img,size=32 --vnc
Regards
Harri
--
Package: amarok
Version: 2.1.1-2
After the upgrade amarok it doesn't play anymore. Esp. the built-in
test produces no beep, neither for the sound card (realtek alc883),
nor for default.
All other players work fine (xine, mplayer, audacious, etc.). Sound
system is Alsa. Of course I have installed
PS: I am not using KDE or Gnome or any other fancy eye-candy. Just plain X
plus fvwm.
Regards
Harri
signature.asc
Description: OpenPGP digital signature
Modestas Vainius wrote:
If it says phonon-backend-gstreamer is installed, purge it and install
phonon-backend-xine instead. Amarok 2 does not seem to work very well with
gstreamer at the moment. What is more, typically gstreamer does not play most
formats out of the box.
You are right,
Package: amarok
Version: 2.1.1-2
Severity: wishlist
Since Amarok2 provides much less functionality (esp there is no Device
support, AFAICS), it would be very nice to allow installing the old
and new Amarok in parallel.
Regards
Harri
signature.asc
Description: OpenPGP digital signature
Package: module-init-tools
Version: 3.9-2
update-modules (called by some poorly maintained packages
at update time) should not sleep for 60 seconds just to
annoy the users. They are innocent.
This nagware-approach puts Debian into a bad light. Please
improve package maintenance instead.
Many
Package: dkms
Version: 2.0.22.0-1
Trying to install drbd8-source I get an error message from
dkms about missing kernel headers (see attachment). Of
course the kernel headers _are_ installed in the usual
location:
# dpkg -S /lib/modules/2.6.30.4/build
raw-kernel-headers-2.6.30.4:
Package: gnupg2
Version: 2.0.9-3.1
Severity: wishlist
Do you think it would be possible for gnupg2 to provide
gnupg? AFAICS gnupg2 provides more configuration options
and is backward compatible to gnupg. That should be sufficient
to resolve the dependencies of other packages, e.g. dpkg-dev
and
Package: qbittorrent
Version: 1.2.0-1
qbittorrent replies at startup:
Could not parse application stylesheet
The usual upload and download speed displays are gone.
Regards
Harri
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL
AFAICS setting the hostname via dhcp still does not work,
even though the debug hooks show that both the old_host_name
and the new_host_name are sent. Attached you can find the
debug output.
Here is a suggested patch:
--- /sbin/dhclient-script.orig 2010-01-09 04:31:48.0 +0100
+++
Package: cryptsetup
Version: 2:1.1.0~rc2-1
I would like to add a key file to an encrypted partition using
device=/dev/sdd1
dd if=/dev/urandom of=/tmp/keyfile bs=256 count=1
cryptsetup luksAddKey ${device} /tmp/keyfile
But cryptsetup asks to enter the (new) passphrase
Attached you can find a patch for the init script. Hope this helps.
Harri
--- /etc/init.d/keepalived.orig 2009-05-28 11:20:36.0 +0200
+++ /etc/init.d/keepalived 2010-01-11 09:52:30.0 +0100
@@ -63,7 +63,7 @@
log_action_begin_msg Restarting $DESC $NAME
Package: virt-manager
Version: 0.8.2-1
Severity: important
To connect to remote kvm hosts via virt-manager I use
qemu+ssh://hdun...@hostname/system
Problem: virt-manager does not terminate the remote helper
applications on exit. After 20 virt-manager sessions I see
20 processes writing
Package: inn2
Version: 2.5.1-1
Tags: patch
I've got an error message from mailpost telling me
-- pipe to |/usr/lib/news/bin/mailpost -b /var/tmp my.newsgroup
generated by my.newsgr...@mailhost --
Use of uninitialized value $_ in pattern match (m//) at
/usr/lib/news/bin/mailpost
Package: qbittorrent
Version: 2.0.7
Seems that #556278 is back: A torrent shows an eta of 6m for more than
an hour now.
Regards
Harri
signature.asc
Description: OpenPGP digital signature
On 01/13/10 23:03, Chris wrote:
Please check the amount of wasted data in main properties panel (lower
part of the window). It happens that there are people sending bad data
and it fails the checking and get registered as wasted data. Please make
sure this is not the case. This is not
Hi R.
Sorry for the delay, but I have good news: The patch seems to work.
In my chroot I got
# dpkg -i cryptmount_4.0.2-1.1_amd64.deb
(Reading database ... 88161 files and directories currently installed.)
Preparing to replace cryptmount 4.0.2-1 (using cryptmount_4.0.2-1.1_amd64.deb)
...
Colin Watson wrote:
UUIDs are properties of filesystems, not disks or partitions, and the
MBR of a hard disk does not have a UUID.
Since MSDOS partitions are limited to 2TByte partitions anyway, it
might be wise to look at alternatives, e.g. GPT. GPT _does_ provide
a UUID in the disk header
After reinstalling SID the problem was gone.
signature.asc
Description: OpenPGP digital signature
Package: qbittorrent
Version: 2.1.3-1
Severity: important
If I select some torrents, and click on delete, then other torrents
than the selected are deleted.
I did not try the web interface, though.
Regards
Harri
signature.asc
Description: OpenPGP digital signature
On 02/09/10 09:01, Chris wrote:
I cannot reproduce here.
When this happens do you have any status or label filter activated on
the left panel?
No.
In transfer list, what column is sorted?
[Upload V] or [Download V].
Just tell me exactly what you did so that I can reproduce.
Looking
On 02/09/10 18:49, Harald Dunkel wrote:
On 02/09/10 09:01, Chris wrote:
In transfer list, what column is sorted?
[Upload V] or [Download V].
PS: It is [Up Speed V] or [Down Speed V], of course.
Regards
Harri
signature.asc
Description: OpenPGP digital signature
Its a shame that this fix didn't made it into testing.
This seriously affects my daily work.
Is version 0.7.5-6 still available online somewhere?
Regards
Harri
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact
Package: virt-manager
Version: 0.8.2-2
Severity: wishlist
If I press ^S (XOFF) on the vnc console terminal, then screen output
stops as expected. If I press ^Q (XON) to continue, then the whole
virt-manager session is stopped, instead of sending it to the login
shell. AFAICS there is no way to
Package: hal
Version: 0.5.14-1
I have granted the user xbmc permission to shutdown, reboot, suspend
and hibernate the PC using
polkit-auth --user xbmc --grant org.freedesktop.hal.power-management.suspend
polkit-auth --user xbmc --grant org.freedesktop.hal.power-management.hibernate
polkit-auth
I would suggest to close this bug. The procedure to build a kernel
module is documented.
Regards
Harri
signature.asc
Description: OpenPGP digital signature
I couldn't reproduce this. I would say this is gone with the recent xorg
update, is it?
Regards
Harri
signature.asc
Description: OpenPGP digital signature
Package: grub-pc
Version: 1.98~20100115-1
Severity: grave
Since the most recent upgrade the PC fails to boot via grub. I get the
menu to select the kernel to boot, but then I see a
Booting 'Debian GNU/Linux, with Linux 2.6.32.3'
error: sparse file not allowed.
Failed to
Sorry, my bug description was not correct: By accident notail was
_not_ set. The workaround mentioned on the Ubuntu bug tracker worked:
https://bugs.launchpad.net/ubuntu/+source/grub2/+bug/464743/comments/6
Suggestion: Maybe the grub2 postinst script could issue a warning for
reiserfs and a
Package: xserver-xorg
Version: 1:7.5+1
To get a left-side Return key I've got a tiny xmodmap file saying
remove Lock = Caps_Lock
keycode 0x42 = Return
It is supposed to redefine the Caps_Lock key.
It worked fine as long as I can remember, but since the most recent
Xserver
Package: console-tools
Version: 1:0.2.3dbs-68
If I install console-tools, then tty6 becomes the default console at
boot time. AFAICS it moves from tty1 to tty6 when entering run level 2.
Since I use some code like
if [ `tty` = /dev/tty1 ]; then
exec startx
fi
in
Package: qbittorent
Version: 2.0.7-1
If I sort the download list by eta, then the idle or paused torrents
are listed first, even though the torrents with a download speed 0
will finish first.
Regards
Harri
signature.asc
Description: OpenPGP digital signature
Package: dclock
Version: 2.2.2-1
The application defaults file for dpkg is not installed correctly.
dpkg shows
# dpkg --contents /var/cache/apt/archives/dclock_2.2.2-1_amd64.deb
drwxr-xr-x root/root 0 2009-02-05 02:21 ./
drwxr-xr-x root/root 0 2009-02-05 02:21 ./etc/
Package: dclock
Version: 2.2.2-1
dclock -fontName 8x13bold
doesn't set the specified font. I would suggest to
make xft support optional. Keep it simple.
Regards
Harri
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble?
I would be interested in this, too.
Esp. for servers running Stable this extended log could
help to document security upgrades.
Regards
Harri
signature.asc
Description: OpenPGP digital signature
Package: initscripts
Version: 2.87dsf-2+b1
If I create /forcefsck and reboot, then the reiserfs
partitions are not verified. fsck -f -a on reiserfs
does not work as you might have expected.
If I run 'fsck -y /dev/sdb3', then reiserfsck does a
read-only test as expected.
I am not sure whether
Package: samba
Version: 2:3.2.5-4lenny6
Even though the current run level doesn't list the
Samba service in /etc/rc2.d, Samba was started during
an upgrade.
This is fatal in my environment (drbd8).
Regards
Harri
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a
Christian Perrier wrote:
What's the output of ls /etc/rc2.d/*samba*?
Also, what's the output of which invoke-rc.d?
# ls /etc/rc2.d/*samba*
ls: cannot access /etc/rc2.d/*samba*: No such file or directory
# which invoke-rc.d
/usr/sbin/invoke-rc.d
Regards
Harri
--
To UNSUBSCRIBE, email
Package: sysv-rc
Version: 2.86.ds1-61
I have disabled several services using insserv -r, e.g.
nfs-common
nfs-kernel-server
samba
They are not supposed to be started or killed by init. Instead they are
under control of heartbeat. The LSB-style headers of these runlevel
Eh, that is not the supported way to disable services. The start
symlinks should be changed to stop symlinks to disable a service the
supported way.
From init's point of view a service can be in one of 3 states in each
runlevel: enabled, disabled or ignored. insserv -r service
moves a service
Petter Reinholdtsen wrote:
[Harald Dunkel]
From init's point of view a service can be in one of 3 states in
each runlevel: enabled, disabled or ignored. insserv -r
service moves a service to the ignored state for all run
levels. For a Linux-HA cluster I need this 3rd state.
Why do you need
Petter Reinholdtsen wrote:
[Harald Dunkel]
I just want to move a service out of reach of init and invoke-rc.d .
Right. And why does it not work to simply disable the service by
replacing start symlinks with stop symlinks in runlevels 2-5? It is
as far as know, the only way to disable
Petter Reinholdtsen wrote:
I belive you are wrong here. If all runlevels have stop symlinks, a
service is never stopped by /etc/init.d/rc because it is assumed to
have happened in the previous runlevel and thus switching to a new
runlevel will not affect the state of the service. Did you
Hi folks,
Do you think it would be possible to get the patch in in
the near future?
Many thanx
Harri
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
AFAICS the new util-linux is stuck in experimental since about
3 weeks. Maybe it would be possible to provide an experimental
hal, too?
Regards
Harri
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact
Package: drbd8
Version: 2:8.0.14-2
Of course I saw the man pages, but I am missing a documentation
package drbd8-doc.
The official home page (http://drbd.org) provides excellent
documentation for the most recent version only (e.g. drbd 8.3.1).
This is fine for Unstable, but insufficient for
Package: nis
Version: 3.17-17
If /etc/passwd contains an empty line, then mknetid (called
by /var/yp/Makefile within a pipe) dies with a core dump.
Testcase:
# (echo; echo +)/tmp/xyz; /usr/lib/yp/mknetid -p /tmp/xyz
Segmentation fault (core dumped)
Regards
Harri
--
To
Regards
Harri
signature.asc
Description: OpenPGP digital signature
Package: passwd
Version: 1:4.1.1-6
If I try to create a new user account using a command
line like
useradd -c John User -b /home -m -g users john
then john becomes a local-only user, even though the host
to run useradd is the master NIS server. The new /etc/passwd
looks like this:
Package: bluez-alsa
Version: 4.40-2
I don't want to complain, but shouldn't the package include a
README in /usr/share/doc/bluez-alsa providing some information
how bluez-alsa is supposed to be configured?
The current READMEs of the bluez packages just provide information
about how to run
Filippo Giunchedi wrote:
Indeed, README.Debian is in the source package but somehow I forgot to include
it in the binary, will do in the next upload.
AFAICS it doesn't provide information about a2dp or alsa, either :-(.
Regards
Harri
signature.asc
Description: OpenPGP digital
Package: bluez-alsa
Version: 4.40-2
Hi folks,
I would like to make the a2dp headset the default audio device on my
laptop (the built-in speakers are just too awful). /etc/asound.conf
says
pcm.!bluetooth {
type bluetooth
device 000:19:7F:6B:9A:04
Filippo Giunchedi wrote:
the current file is this:
http://svn.debian.org/wsvn/pkg-bluetooth/packages/bluez/trunk/debian/README.Debian
patches are welcome to improve said file
The new README is surely much better wrt Alsa and a2dp.
I would suggest to keep the section about
Filippo Giunchedi wrote:
On Sat, Jun 06, 2009 at 06:51:21PM +0200, Filippo Giunchedi wrote:
I'd rather not play with the storage itself, the dbus API provides means for
bonding devices thus I'll try to document that.
I'd like if you can try this:
dbus-send --system --print-reply
Filippo Giunchedi wrote:
On Sun, Jun 07, 2009 at 07:12:52AM +0200, Harald Dunkel wrote:
BTW, bluez-alsa has a very short list of dependencies. Esp. there
is no dbus. Is this correct?
yes, bluez-alsa is only the alsa part, the bluez part is already included in
bluez package
bluez-alsa
Package: nfs-kernel-server
Version: 1:1.1.2-6lenny1
/etc/init.d/nfs-kernel-server doesn't stop the nfs daemons.
This might be a problem in a HA environment. Usually the
resources are supposed to go away before the IP address is
bound to another host.
Regards
Harri
--
To UNSUBSCRIBE,
PS: Seems that this _is_ a fatal problem for heartbeat.
/etc/init.d/nfs-kernel-server stop (called by heartbeat)
seems to use SIGINT to kill the nfs daemons, but they
don't go away. This keeps the local mount point of an
exported disk busy, and after 2 minutes waiting heartbeat
pulls the plug and
Nicolas François wrote:
Did it used to work previously (with which version)?
Sorry, but I have no chance to try.
Does adduser behave correctly regarding NIS?
adduser calls useradd to do the real work. Same problem.
Is it legitimate in some cases to add an user locally, and not to NIS.
Package: dbus
Version: 1.2.16-1
Trying to update dbus I got this:
Preparing to replace libdbus-1-dev 1.2.14-3 (using
.../libdbus-1-dev_1.2.16-1_amd64.deb) ...
Unpacking replacement libdbus-1-dev ...
Preparing to replace libdbus-1-3 1.2.14-3 (using
.../libdbus-1-3_1.2.16-1_amd64.deb) ...
Package: heartbeat
Version: 2.1.3-6lenny1
Tags: patch
Usually Heartbeat uses a drbd partition to share configuration
and service state information between the 2 hosts. Drbd itself
is not started by Heartbeat, so /etc/init.d/heartbeat should
show:
% diff -u heartbeat.orig heartbeat
---
Package: nagios-nrpe-server
Version: 2.12-1
If root does /etc/init.d/nagios-nrpe-server restart or
invoke-rc.d nagios-nrpe-server restart, then nrpe runs
with a new environment inherited from root's interactive
shell. DISPLAY, LANG, PATH and all the others are inherited
by the plugins. This has
Simon Horman wrote:
show:
Hi,
I assume this would have no effect on systems where
drdb isn't installed. Is that correct?
Yes, I tried it just this minute.
This change doesn't affect systems _with_ drbd installed,
either. In the static runlevel configuration (without insserv)
drbd is
Package: udev
Version: 0.125-7+lenny1
Even though /etc/udev/rules.d/91-permissions.rules says
KERNEL==tun, MODE=0666
the mode of /dev/net/tun is set to 0600:
# ls -al /dev/net/tun
crw--- 1 root root 10, 200 Jul 2 15:42 /dev/net/tun
# grep -r tun /etc/udev/rules.d/*
PS: Of course I made sure that the udev *.rules files included
with the initrd are in sync with /etc/udev/rules.d. The initrd
was rebuilt after changing /etc/udev/rules.d.
Regards
Harri
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe.
Please reopen. Reason:
I do not see that any other package is involved. AFAICS /dev/net/tun
is created by /lib/udev/create_static_nodes, reading a list of
/dev entries from /etc/udev/links.conf. It ignores the regular udev
rules completely. This is surely not what one would expect.
If there is
Sorry for my ignorance, but maybe I missed something here.
Why is /dev/net/tun created by /lib/udev/create_static_nodes?
Regards
Harri
signature.asc
Description: OpenPGP digital signature
Regards
Harri
signature.asc
Description: OpenPGP digital signature
Running modprobe tun does.
???
Regards
Harri
signature.asc
Description: OpenPGP digital signature
Marco d'Itri wrote:
On Jul 05, Harald Dunkel harald.dun...@t-online.de wrote:
Why is /dev/net/tun created by /lib/udev/create_static_nodes?
Because nothing else loads the module.
AFAICS creating /dev/net/tun doesn't trigger loading the
module either. Even if it would, the tun module
Marco d'Itri wrote:
On Jul 05, Harald Dunkel harald.dun...@t-online.de wrote:
What would you suggest?
echo tun /etc/modules
I would suggest to remove the tun entry from /etc/udev/links.conf.
This would make sure that there is no time slot with /dev/net/tun
having unwanted access
Package: smartmontools
Version: 5.38-3
Severity: wishlist
I would suggest to make -d removable the default (see smartd.conf(5)).
One missing disk should not affect watching the others, even if it is
not intended to be removed. Sending an EMail about the missing disk is
surely OK.
Regards
Harri
301 - 400 of 1828 matches
Mail list logo