gmane.linux.debian.devel.sysvinit

Bug#706676: sysvinit: /sbin/init doesn'tsupport LXC gracefull shutdown via lxc-shutdown

Pallai Roland — 2013-06-15T22:59:53

The problem is lxc has to watch the jail's utmp file to guess when it
wants to halt, because Linux kernels before version 3.4 does not
provide any function to reboot or halt a container from inside. It has
leaded to a dirty hack in lxc-utils that failed now.

First problem is the overlaid /run; the Wheezy initscript mounts a
tmpfs in /run and (re)creates utmp file there but lxc-utils won't
notice it. This could be fixed several ways:
1. do not give permission for the container to mount file systems,
drop sys_admin capability
2. mount tmpfs in /run from the container's lxc config before the
container mounts it

Second problem is the absolute symlink in /var/run; lxc-utils does not
expect absolute symlink there so it will watch the host's utmp file
for changes that pretty useless. You cannot replace this absolute link
in a Wheezy container, the initscript will restore that on next boot.
A trivial solution to patch lxc-utils to prefer /run over /var/run for
utmp watching, see my attachment.

Yes, my patch just

Bug#712425: initscripts: Document how todisable "concurrency" boot

Camaleón — 2013-06-15T20:53:15

Package: initscripts
Version: 2.88dsf-41
Severity: normal

Running Wheezy, I noticed one of my services (from a third party vendor) 
was not started when system booted. All seemed to be fine (LSB headers, 
required services, links under defined runlevels...). The service (an 
utility to monitor UPS units) could be started manually ("service upsmon 
start") with no problem at all.

By digging a bit more into this, I remembered "concurrency" was the 
default since squeeze and it could be disable by setting 
"CONCURRENY=none" in "/etc/default/rcS" but this seems to be no more an 
option ("man rcS" neither mentions this variable).

So I wondered how could I disable concurrency. 

The answer was under "/etc/init.d/rc" script itself by creating an empty 
file "/etc/init.d/.legacy-bootordering" that sets CONCURRENY variable to 
"none") but I don't think is a convenient place to document this option.

P.S. The above mentioned service was indeed unable to start with 
concurrency "on" but worked fine with the old sequ

Bug#706676: sysvinit: /sbin/init doesn'tsupport LXC gracefull shutdown via lxc-shutdown

Ivan Vilata i Balaguer — 2013-06-13T11:03:17

Package: sysvinit
Version: 2.88dsf-41
Followup-For: Bug #706676

I just discovered that not dropping the ``sys_boot`` capability from a
container fixes the issue, plus it allows the container to properly reboot
(instead of launching a bare shell in the case of Debian containers).

So maybe it was the old kernel which handled that capability in an incorrect
way and the problem is not in sysvinit at all.

Processed (with 1 errors): Re: Bug#712078: Acknowledgement (btrfs-progs: fsck.btrfs -a is not working)

Debian Bug Tracking System — 2013-06-13T08:33:12

Processing commands for control< at >bugs.debian.org:

Bug #712078 [btrfs-tools] btrfs-progs: fsck.btrfs -a is not working
Unable to merge bugs because:
affects of #701956 is 'btrfs-tools' not ''
package of #701956 is 'initscripts' not 'btrfs-tools'
Failed to merge 712078: Did not alter merged bugs
Debbugs::Control::set_merged('transcript', 'GLOB(0x2c7b918)', 'requester', 'Sebastian Andrzej Siewior <sebastian< at >breakpoint.cc>', 'request_addr', 'control< at >bugs.debian.org', 'request_msgid', '<20130613081914.GA2355< at >breakpoint.cc>', 'request_subject', ...) called at /usr/local/lib/site_perl/Debbugs/Control/Service.pm line 538
eval {...} called at /usr/local/lib/site_perl/Debbugs/Control/Service.pm line 537
Debbugs::Control::Service::control_line('line', undef, 'clonebugs', 'HASH(0x2bbcae8)', 'limit', 'HASH(0x2bbc4d0)', 'common_control_options', 'ARRAY(0x2bbc518)', 'errors', ...) called at /usr/lib/debbugs/service line 474

Stopping processing here.

Please contact me if you need assistance.

Bug#712023: initscripts: wait for child toexit during shutdown/reboot

erik — 2013-06-12T08:55:53

Package: initscripts
Version: 2.88dsf-41
Severity: important

Dear Maintainer,
*** Please consider answering these questions, where appropriate ***

   * What led up to the situation?
I use virtualbox. I have virtualmachines on my box and want to bring them down
or save the state there in, before the system is going down.
   * What exactly did you do (or not do) that was effective (or
     ineffective)?
I have made a script to readout the vm's that are running and save there state
one by one. This takes some time. At the beginning no vm was saved right. I did
some tunning with the scripts and have now 5-10 sec's before the other
vboxservices are terminated. So i can bring 3 maybe 4 syetems the rightway
down.
I did look in the documentation, and i didnot found a usefull option , that
could ensure me that the vm is brought down the rightway. (wait for child to
terminated)

   * What was the outcome of this action?

   * What outcome did you expect instead?
I am looking for a manner that the system will wait fo

Bug#706676: sysvinit: /sbin/init doesn'tsupport LXC gracefull shutdown via lxc-shutdown

Ivan Vilata i Balaguer — 2013-06-11T14:19:45

Package: sysvinit
Version: 2.88dsf-41
Followup-For: Bug #706676

The bug still exists with kernel 3.9-1-amd64 (3.9.5-1).

Bug#711963: [initscripts] does not mount /tmpas tmpfs

Marek Straka — 2013-06-11T15:33:26

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Package: initscripts
Version: 2.88dsf-41
Severity: normal

- --- Please enter the report below this line. ---

Just edited /etc/default/tmpfs to:

# Configuration for tmpfs filesystems mounted in early boot, before
# filesystems from /etc/fstab are mounted.  For information about
# these variables see the tmpfs(5) manual page.

# /run is always mounted as a tmpfs on systems which support tmpfs
# mounts.

# mount /run/lock as a tmpfs (separately from /run).  Defaults to yes;
# set to no to disable (/run/lock will then be part of the /run tmpfs,
# if available).
#RAMLOCK=yes

# mount /run/shm as a tmpfs (separately from /run).  Defaults to yes;
# set to no to disable (/run/shm will then be part of the /run tmpfs,
# if available).
#RAMSHM=yes

# mount /tmp as a tmpfs.  Defaults to no; set to yes to enable (/tmp
# will be part of the root filesystem if disabled).  /tmp may also be
# configured to be a separate mount in /etc/fstab.
RAMTMP=yes

# Size limits.  Please

Bug#436703: #436703: - patch broken - only 8scsi devices?

Bernd Zeimetz — 2007-09-21T08:49:37

Hi,

I'm documenting a discussion about this bug on #debian.de, which had the
following results:

++#define MAX_SCSI_DISKS 8

unfortunately that's broken. It's pretty easy to have way more than 8
scsi devices before the drives in your machine are initialized. You'll
have to take care of all scsi devices.


++"/sys/block/sdX/device/scsi_disk:*/manage_start_stop";

libsysfs should provide highlevel functions to access this.

Bug#436703: Accepted linux-2.62.6.22-4~bpo40+1 (source all amd64)

Norbert Tretkowski — 2007-09-21T07:41:37

Am Freitag, den 21.09.2007, 09:25 +0200 schrieb Jan Wagner:

I'm not aware of any workaround for this problem. The perfect solution
is to update sysvinit in stable.

Norbert




_______________________________________________
Pkg-sysvinit-devel mailing list
Pkg-sysvinit-devel< at >lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-sysvinit-devel

Re: Stable update for sysvinit

Roger Leigh — 2013-06-10T12:37:39

Hi Adam,

Sorry for the delay here, I've been in hospital for a few days and
moving house immediately after, leaving me currently in a poorly
mobile state and lacking internet.  Currently using a work computer
but no means to build packages for upload or sign them here.  I should
have internet at home by next weekend, but I'm not really in a
position to be able to do the work until that point in any case.

It's all here:
http://anonscm.debian.org/gitweb/?p=collab-maint/sysvinit;a=shortlog;h=refs/heads/wheezy
it just needs building and uploading with any changes to the version
number you need.


Regards,
Roger

Bug#698751: Bug#698751: bootlogd does not work

Kyle Guilbert — 2013-06-06T16:07:14

I retract my statement. It does work now, after doing another install. My
previous install was slightly nonstandard which must have caused the
problem.

By the way, /etc/default/bootlogd neither exists in the new 'bootlogd'
package nor does it appear to be necessary.

On Thu, Jun 6, 2013 at 12:49 AM, Petter Reinholdtsen <pere< at >hungry.com>wrote:

Bug#698751: Bug#698751: bootlogd does not work

Petter Reinholdtsen — 2013-06-06T04:49:47

[Kyle Guilbert]

Even after enabling it in /etc/default/bootlogd?

Bug#698751: bootlogd does not work

Kyle Guilbert — 2013-06-06T03:49:31

I concur that bootlogd is not functional, on a fresh install of Debian 7. My /var/log/boot simply says "(Nothing has been logged yet.)". Bootlogd version tested is 2.88dsf-41.

Bug#696910: initscripts: cannot shut down when/ on network - iscsi

Ferenc Wagner — 2013-06-04T14:14:18

Hi,

I was also bitten by this issue; Dear Maintainer, please consider fixing
it!  I applied to following patches to the reboot and halt scripts,
based on similar checks in the networking and open-iscsi init scripts:

--- /etc/init.d/reboot.orig2012-11-20 12:03:28.737794685 +0100
+++ /etc/init.d/reboot2012-11-20 12:07:20.209804174 +0100
< at >< at > -17,7 +17,11 < at >< at >
 # Message should end with a newline since kFreeBSD may
 # print more stuff (see #323749)
 log_action_msg "Will now restart"
-reboot -d -f -i
+netdown="-i"
+if [ -e /etc/iscsi/iscsi.initramfs ]; then
+netdown=""
+fi
+reboot -d -f $netdown
 }
 
 case "$1" in
--- /etc/init.d/halt.orig2012-11-20 12:03:25.237794541 +0100
+++ /etc/init.d/halt2012-11-20 12:05:23.209799378 +0100
< at >< at > -53,9 +53,9 < at >< at >
 fi
 
 # Make it possible to not shut down network interfaces,
-# needed to use wake-on-lan
+# needed to use wake-on-lan or to synchronize the iSCSI cache
 netdown="-i"
-if [ "$NETDOWN" = "no" ]; then
+if [ "$NETDOWN" = "no" ] || [ -e /etc/iscsi/iscsi

Processed: block 701559 with 710571, block 710569 with 710571, block 710559 with 710571

Debian Bug Tracking System — 2013-06-03T06:36:12

Processing commands for control< at >bugs.debian.org:

Bug #701559 {Done: Vincent Legout <vlegout< at >debian.org>} [xboard] Please update xboard to stable version (4.7.0)
701559 was blocked by: 710569
701559 was not blocking any bugs.
Added blocking bug(s) of 701559: 683654 and 710571
Bug #710569 [initscripts] Update /etc/init.d/skeleton to stop checking $VERBOSE before calling log_* functions
710569 was not blocked by any bugs.
710569 was blocking: 701559 710559
Added blocking bug(s) of 710569: 683654 and 710571
Bug #710559 [lintian] lintian: init.d-script-call-internal-API should document replacement (remove checks for $VERBOSE and use log_* functions)
710559 was blocked by: 710569
710559 was not blocking any bugs.
Added blocking bug(s) of 710559: 683654 and 710571
Stopping processing here.

Please contact me if you need assistance.

Processed: block 710559 with 710569

Debian Bug Tracking System — 2013-06-01T00:33:14

Processing commands for control< at >bugs.debian.org:

Bug #710559 [lintian] lintian: init.d-script-call-internal-API should document replacement (remove checks for $VERBOSE and use log_* functions)
710559 was not blocked by any bugs.
710559 was not blocking any bugs.
Added blocking bug(s) of 710559: 710569
Stopping processing here.

Please contact me if you need assistance.

Processed: tagging 710569

Debian Bug Tracking System — 2013-05-31T23:09:22

Processing commands for control< at >bugs.debian.org:

Bug #710569 [initscripts] Update /etc/init.d/skeleton to stop checking $VERBOSE before calling log_* functions
Ignoring request to alter tags of bug #710569 to the same tags previously set
Stopping processing here.

Please contact me if you need assistance.

Processed: Patch for /etc/init.d/skeleton;more bug twiddling

Debian Bug Tracking System — 2013-05-31T23:09:19

Processing commands for control< at >bugs.debian.org:

Bug #710569 [initscripts] initscripts: Update /lib/lsb/init-functions log_* to check $VERBOSE, and update /etc/init.d/skeleton accordingly
Added tag(s) patch.
Bug #701559 {Done: Vincent Legout <vlegout< at >debian.org>} [xboard] Please update xboard to stable version (4.7.0)
701559 was not blocked by any bugs.
701559 was not blocking any bugs.
Added blocking bug(s) of 701559: 710569
Bug #710569 [initscripts] initscripts: Update /lib/lsb/init-functions log_* to check $VERBOSE, and update /etc/init.d/skeleton accordingly
Bug 710569 cloned as bug 710571
710571 was not blocked by any bugs.
710571 was blocking: 701559
Added blocking bug(s) of 710571: 701559
Bug #710571 [initscripts] initscripts: Update /lib/lsb/init-functions log_* to check $VERBOSE, and update /etc/init.d/skeleton accordingly
Bug reassigned from package 'initscripts' to 'lsb-base'.
Ignoring request to alter found versions of bug #710571 to the same values previously set
Ignoring request to alter fixed

Bug#710569: Patch for /etc/init.d/skeleton;more bug twiddling

Josh Triplett — 2013-05-31T23:05:58

tags 710569 + patch
block 701559 by 710569
clone 710569 -1
reassign -1 lsb-base
retitle 710569 Update /etc/init.d/skeleton to stop checking $VERBOSE before calling log_* functions
retitle -1 Update /lib/lsb/init-functions log_* to check $VERBOSE
block 710569 by -1
thanks

Sorry, should have made that two separate bugs, one for skeleton (in
initscripts) and one for init-functions (in lsb-base).  I'll submit a
patch for the latter in a moment.

The attached git patch fixes skeleton in initscripts.  As the above
blocking indicates, though, it shouldn't go in before the corresponding
change to init-functions in lsb-base.

Also, the lintian change for 710569 should state the need for a
versioned dependency on whatever version of lsb-base incorporates this
change.

- Josh Triplett

Processed: Re: Bug#710559: lintian: init.d-script-call-internal-API contradicts /etc/init.d/skeleton

Debian Bug Tracking System — 2013-05-31T22:45:27

Processing commands for control< at >bugs.debian.org:

Bug #710559 [lintian] lintian: init.d-script-call-internal-API contradicts /etc/init.d/skeleton
Changed Bug title to 'lintian: init.d-script-call-internal-API should document replacement (remove checks for $VERBOSE and use log_* functions)' from 'lintian: init.d-script-call-internal-API contradicts /etc/init.d/skeleton'
Bug #710559 [lintian] lintian: init.d-script-call-internal-API should document replacement (remove checks for $VERBOSE and use log_* functions)
Bug 710559 cloned as bug 710569
Bug #710569 [lintian] lintian: init.d-script-call-internal-API should document replacement (remove checks for $VERBOSE and use log_* functions)
Bug reassigned from package 'lintian' to 'initscripts'.
No longer marked as found in versions lintian/2.5.13.
Ignoring request to alter fixed versions of bug #710569 to the same values previously set
Bug #710569 [initscripts] lintian: init.d-script-call-internal-API should document replacement (remove checks for $VERBOSE and use

Bug#710519: /etc/init.d/skeleton: skeletonfile uses internal API

Laurent Bigonville — 2013-05-31T15:19:33

Package: initscripts
Version: 2.88dsf-41
Severity: normal
File: /etc/init.d/skeleton

Hi,

Lintian has just added a new "init.d-script-call-internal-API" check
which is triggered when a file inside the "/lib/init/" directory is
sourced.

Removing the sourcing of "/lib/init/vars.sh" would also require to
remove the check of the VERBOSE variable.

Cheers

Laurent Bigonville