Bug#832336: jessie-pu: package systemd/215-17+deb8u5

[Adam D. Barratt]

Control: tags -1 + pending

On Sat, 2016-09-03 at 00:09 +0200, Michael Biebl wrote:
> Am 02.09.2016 um 22:12 schrieb Adam D. Barratt:
> > Sorry for the delays. I've been arguing with myself about the fsck fix
> > as well, but I'd be happy for the remainder to be uploaded.
> 
> Ok then, uploaded with that one particular change backed out.

Thanks; flagged for acceptance into p-u.

Regards,

Adam

Bug#832336: jessie-pu: package systemd/215-17+deb8u5

[Michael Biebl]

Am 02.09.2016 um 22:12 schrieb Adam D. Barratt:
> Sorry for the delays. I've been arguing with myself about the fsck fix
> as well, but I'd be happy for the remainder to be uploaded.

Ok then, uploaded with that one particular change backed out.

Regards,
Michael

-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?



signature.asc
Description: OpenPGP digital signature

Bug#832336: jessie-pu: package systemd/215-17+deb8u5

[Adam D. Barratt]

# see note at end
Control: tags -1 + confirmed

On Wed, 2016-08-31 at 01:43 +0200, Michael Biebl wrote:
> Am 18.08.2016 um 23:41 schrieb Michael Biebl:
> > Am 04.08.2016 um 22:13 schrieb Michael Biebl:
> >> Am 03.08.2016 um 22:41 schrieb Julien Cristau:
> >>> On Wed, Aug  3, 2016 at 22:32:16 +0200, Michael Biebl wrote:
> >>>
>  Am 29.07.2016 um 23:00 schrieb Michael Biebl:
> > Am 29.07.2016 um 14:00 schrieb Julien Cristau:
> >> This looks fine in general, I'm just wondering if the effect of bumping
> >> the util-linux dependency on upgrades from wheezy was 
> >> considered/tested?
> >
> > It was not specifically tested since I didn't expect any breakage
> > because of it. But I did it now for a basic wheezy system the upgrade to
> > jessie (with the updated systemd packages) worked fine. The tightened
> > dep on u-l seems to not cause any issues.
> >
> > I've attached the typescript from the wheezy → jessie dist-upgrade.
> 
>  Is that sufficient or do you still have concerns?
> 
> >>> I'm definitely concerned about changing dependencies in the base system
> >>> in stable, yes.  Is the parallel fsck bug that bad?
> >>
> >> I would really like to include that fsck fix. Parallel fsck on rotating
> >> disks is bad. It degrades the performance and we had reports from
> >> concerned users, that their disks made very strange noises when the
> >> heads where constantly repositioned.
> >> This doesn't affect SSDs, but I guess HDD are still very common on a
> >> typical server.
> > 
> > With the 8.6 release being close and no further response, I assume you
> > still object to this change. I would still like to see the other fixes
> > land in 8.6, so if you are ok with the other changes, can I proceed with
> > the upload if I exclude the 'fsck -l' fix?
> 
> 
> Hi, would be great to have a final decisions on this issue so I can
> proceed with the upload. I'd hate to see the accumulated fixes for
> systemd not to land in 8.6.

Sorry for the delays. I've been arguing with myself about the fsck fix
as well, but I'd be happy for the remainder to be uploaded.

Regards,

Adam

Bug#832336: jessie-pu: package systemd/215-17+deb8u5

[Michael Biebl]

Am 18.08.2016 um 23:41 schrieb Michael Biebl:
> Am 04.08.2016 um 22:13 schrieb Michael Biebl:
>> Am 03.08.2016 um 22:41 schrieb Julien Cristau:
>>> On Wed, Aug  3, 2016 at 22:32:16 +0200, Michael Biebl wrote:
>>>
 Am 29.07.2016 um 23:00 schrieb Michael Biebl:
> Am 29.07.2016 um 14:00 schrieb Julien Cristau:
>> This looks fine in general, I'm just wondering if the effect of bumping
>> the util-linux dependency on upgrades from wheezy was considered/tested?
>
> It was not specifically tested since I didn't expect any breakage
> because of it. But I did it now for a basic wheezy system the upgrade to
> jessie (with the updated systemd packages) worked fine. The tightened
> dep on u-l seems to not cause any issues.
>
> I've attached the typescript from the wheezy → jessie dist-upgrade.

 Is that sufficient or do you still have concerns?

>>> I'm definitely concerned about changing dependencies in the base system
>>> in stable, yes.  Is the parallel fsck bug that bad?
>>
>> I would really like to include that fsck fix. Parallel fsck on rotating
>> disks is bad. It degrades the performance and we had reports from
>> concerned users, that their disks made very strange noises when the
>> heads where constantly repositioned.
>> This doesn't affect SSDs, but I guess HDD are still very common on a
>> typical server.
> 
> With the 8.6 release being close and no further response, I assume you
> still object to this change. I would still like to see the other fixes
> land in 8.6, so if you are ok with the other changes, can I proceed with
> the upload if I exclude the 'fsck -l' fix?


Hi, would be great to have a final decisions on this issue so I can
proceed with the upload. I'd hate to see the accumulated fixes for
systemd not to land in 8.6.

Regards,
Michael

-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?



signature.asc
Description: OpenPGP digital signature

Bug#832336: jessie-pu: package systemd/215-17+deb8u5

[Michael Biebl]

Am 04.08.2016 um 22:13 schrieb Michael Biebl:
> Am 03.08.2016 um 22:41 schrieb Julien Cristau:
>> On Wed, Aug  3, 2016 at 22:32:16 +0200, Michael Biebl wrote:
>>
>>> Am 29.07.2016 um 23:00 schrieb Michael Biebl:
 Am 29.07.2016 um 14:00 schrieb Julien Cristau:
> This looks fine in general, I'm just wondering if the effect of bumping
> the util-linux dependency on upgrades from wheezy was considered/tested?

 It was not specifically tested since I didn't expect any breakage
 because of it. But I did it now for a basic wheezy system the upgrade to
 jessie (with the updated systemd packages) worked fine. The tightened
 dep on u-l seems to not cause any issues.

 I've attached the typescript from the wheezy → jessie dist-upgrade.
>>>
>>> Is that sufficient or do you still have concerns?
>>>
>> I'm definitely concerned about changing dependencies in the base system
>> in stable, yes.  Is the parallel fsck bug that bad?
> 
> I would really like to include that fsck fix. Parallel fsck on rotating
> disks is bad. It degrades the performance and we had reports from
> concerned users, that their disks made very strange noises when the
> heads where constantly repositioned.
> This doesn't affect SSDs, but I guess HDD are still very common on a
> typical server.

With the 8.6 release being close and no further response, I assume you
still object to this change. I would still like to see the other fixes
land in 8.6, so if you are ok with the other changes, can I proceed with
the upload if I exclude the 'fsck -l' fix?

Michael



-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?



signature.asc
Description: OpenPGP digital signature

Bug#832336: jessie-pu: package systemd/215-17+deb8u5

[Michael Biebl]

Am 03.08.2016 um 22:41 schrieb Julien Cristau:
> On Wed, Aug  3, 2016 at 22:32:16 +0200, Michael Biebl wrote:
> 
>> Am 29.07.2016 um 23:00 schrieb Michael Biebl:
>>> Am 29.07.2016 um 14:00 schrieb Julien Cristau:
 This looks fine in general, I'm just wondering if the effect of bumping
 the util-linux dependency on upgrades from wheezy was considered/tested?
>>>
>>> It was not specifically tested since I didn't expect any breakage
>>> because of it. But I did it now for a basic wheezy system the upgrade to
>>> jessie (with the updated systemd packages) worked fine. The tightened
>>> dep on u-l seems to not cause any issues.
>>>
>>> I've attached the typescript from the wheezy → jessie dist-upgrade.
>>
>> Is that sufficient or do you still have concerns?
>>
> I'm definitely concerned about changing dependencies in the base system
> in stable, yes.  Is the parallel fsck bug that bad?

I would really like to include that fsck fix. Parallel fsck on rotating
disks is bad. It degrades the performance and we had reports from
concerned users, that their disks made very strange noises when the
heads where constantly repositioned.
This doesn't affect SSDs, but I guess HDD are still very common on a
typical server.

Regards,
Michael


-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?



signature.asc
Description: OpenPGP digital signature

Bug#832336: jessie-pu: package systemd/215-17+deb8u5

[Julien Cristau]

On Wed, Aug  3, 2016 at 22:32:16 +0200, Michael Biebl wrote:

> Am 29.07.2016 um 23:00 schrieb Michael Biebl:
> > Am 29.07.2016 um 14:00 schrieb Julien Cristau:
> >> This looks fine in general, I'm just wondering if the effect of bumping
> >> the util-linux dependency on upgrades from wheezy was considered/tested?
> > 
> > It was not specifically tested since I didn't expect any breakage
> > because of it. But I did it now for a basic wheezy system the upgrade to
> > jessie (with the updated systemd packages) worked fine. The tightened
> > dep on u-l seems to not cause any issues.
> > 
> > I've attached the typescript from the wheezy → jessie dist-upgrade.
> 
> Is that sufficient or do you still have concerns?
> 
I'm definitely concerned about changing dependencies in the base system
in stable, yes.  Is the parallel fsck bug that bad?

Cheers,
Julien

Bug#832336: jessie-pu: package systemd/215-17+deb8u5

[Michael Biebl]

Am 29.07.2016 um 23:00 schrieb Michael Biebl:
> Am 29.07.2016 um 14:00 schrieb Julien Cristau:
>> This looks fine in general, I'm just wondering if the effect of bumping
>> the util-linux dependency on upgrades from wheezy was considered/tested?
> 
> It was not specifically tested since I didn't expect any breakage
> because of it. But I did it now for a basic wheezy system the upgrade to
> jessie (with the updated systemd packages) worked fine. The tightened
> dep on u-l seems to not cause any issues.
> 
> I've attached the typescript from the wheezy → jessie dist-upgrade.

Is that sufficient or do you still have concerns?

Michael

-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?



signature.asc
Description: OpenPGP digital signature

Bug#832336: jessie-pu: package systemd/215-17+deb8u5

[Michael Biebl]

Am 29.07.2016 um 14:00 schrieb Julien Cristau:
> This looks fine in general, I'm just wondering if the effect of bumping
> the util-linux dependency on upgrades from wheezy was considered/tested?

It was not specifically tested since I didn't expect any breakage
because of it. But I did it now for a basic wheezy system the upgrade to
jessie (with the updated systemd packages) worked fine. The tightened
dep on u-l seems to not cause any issues.

I've attached the typescript from the wheezy → jessie dist-upgrade.

Regards,
Michael

-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?


typescript.gz
Description: application/gzip


signature.asc
Description: OpenPGP digital signature

Bug#832336: jessie-pu: package systemd/215-17+deb8u5

[Julien Cristau]

On Sun, Jul 24, 2016 at 18:59:04 +0200, Michael Biebl wrote:

> Am 24.07.2016 um 14:12 schrieb Michael Biebl:
> > for the upcoming jessie stable release, I'd like to update the systemd
> > package with the following fixes:
> 
> [..]
> 
> > The complete debdiff is attached as well.
> 
> I would like to include one more fix. The request for that came up just
> today, see [1]. Sorry for the bad timing.
> 
> It's an upstream commit [2] which is part of v230 and ensures correct
> shutdown order and kills the user sessions before the network stack is
> stopped.
> This is needed for open SSH connections to be terminated cleanly.
> 
> For this commit to apply cleanly, it also requires upstream commit [3],
> which seems like a reasonable change as well.
> 
> An updated debdiff is attached.
> 
> Please let me know if you are ok with this updated debdiff.
> 
Hi,

This looks fine in general, I'm just wondering if the effect of bumping
the util-linux dependency on upgrades from wheezy was considered/tested?

Cheers,
Julien

Bug#832336: jessie-pu: package systemd/215-17+deb8u5

[Michael Biebl]

Am 24.07.2016 um 14:12 schrieb Michael Biebl:
>   * getty-static.service: Only start if we have a working VC subsystem.
> Use ConditionPathExists=/dev/tty0, the same check as in getty@.service,
> to determine whether we have a functional VC subsystem and we should
> start any gettys. (Closes: #824779)
> 
> http://anonscm.debian.org/cgit/pkg-systemd/systemd.git/commit/?h=jessie=95dc10016
> 
> getty-static.service is a Debian specific addition, so this change is
> not an upstream cherry-pick. This fix hasn't been uploaded to unstable
> yet, but is atm only committed to master and pending. So strictly
> speaking it violates the stable upload policy.

This change has now been uploaded to unstable as well.

Regards,
Michael

-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?



signature.asc
Description: OpenPGP digital signature

Bug#832336: jessie-pu: package systemd/215-17+deb8u5

[Michael Biebl]

Am 24.07.2016 um 14:12 schrieb Michael Biebl:
> Package: release.debian.org
> Severity: normal
> Tags: jessie
> User: release.debian@packages.debian.org
> Usertags: pu
> 
> Dear release team,
> 
> for the upcoming jessie stable release, I'd like to update the systemd
> package with the following fixes:

[..]

> The complete debdiff is attached as well.

I would like to include one more fix. The request for that came up just
today, see [1]. Sorry for the bad timing.

It's an upstream commit [2] which is part of v230 and ensures correct
shutdown order and kills the user sessions before the network stack is
stopped.
This is needed for open SSH connections to be terminated cleanly.

For this commit to apply cleanly, it also requires upstream commit [3],
which seems like a reasonable change as well.

An updated debdiff is attached.

Please let me know if you are ok with this updated debdiff.

Regards,
Michael



[1] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=832155#50
[2]
https://github.com/systemd/systemd/commit/8c856804780681e135d98ca94d08afe247557770
[3]
https://github.com/systemd/systemd/commit/efb3e19be9c568974b221990b9e84fb5304c5537
-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?
diff --git a/debian/changelog b/debian/changelog
index 974bbb0..3f0f176 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,3 +1,31 @@
+systemd (215-17+deb8u5) stable; urgency=medium
+
+  * Use the right timeout for stop processes we fork. This ensures that
+services are properly killed after a given timeout. (Closes: #813702)
+  * Re-enable fsck -l. This avoids that systemd-fsck runs fsck in parallel for
+file systems on the same disk. Parallel fsck can lead to suboptimal
+behaviour on rotational hard disks. This requires util-linux 2.25, so bump
+the dependency accordingly. (Closes: #824963)
+  * Don't reset log level to NOTICE if we get quiet on the kernel cmdline.
+(Closes: #828006)
+  * Fix prepare priority queue comparison function in sd-event.
+Otherwise a disabled event source can get swapped with an enabled one
+and cause a severe sd-event malfunction, breaking the event loop.
+(Closes: #789796)
+  * Update links to kernel.org cgroup documentation.
+The systemd.resource-control man page had references to /cgroups/ which
+moved to /cgroup-v1/. (Closes: #819970)
+  * Don't start console-getty.service when /dev/console is missing.
+Avoids repeated unsuccessful start attempts of agetty inside (docker)
+containers. (Closes: #829537)
+  * Order systemd-user-sessions.service after nss-user-lookup.target. We
+should not allow logins before NIS/LDAP users are available.
+  * Order systemd-user-sessions.service after network.target. That way we can
+be sure that local users are logged out and SSH sessions are ended cleanly
+before the network is shut down when the system goes down.
+
+ -- Michael Biebl   Sun, 24 Jul 2016 18:55:54 +0200
+
 systemd (215-17+deb8u4) stable; urgency=medium
 
   [ Martin Pitt ]
diff --git a/debian/control b/debian/control
index aa02024..cdcc5a4 100644
--- a/debian/control
+++ b/debian/control
@@ -51,7 +51,7 @@ Pre-Depends: ${shlibs:Pre-Depends},
 Depends: ${shlibs:Depends},
  ${misc:Depends},
  libsystemd0 (= ${binary:Version}),
- util-linux (>= 2.19.1-2),
+ util-linux (>= 2.25),
  mount (>= 2.21),
  initscripts (>= 2.88dsf-53.2),
  sysv-rc,
diff --git a/debian/extra/getty-static.service 
b/debian/extra/getty-static.service
index 8636be5..0a1203f 100644
--- a/debian/extra/getty-static.service
+++ b/debian/extra/getty-static.service
@@ -1,6 +1,6 @@
 [Unit]
 Description=getty on tty2-tty6 if dbus and logind are not available
-ConditionPathExists=/dev/tty2
+ConditionPathExists=/dev/tty0
 ConditionPathExists=!/lib/systemd/system/dbus.service
 
 [Service]
diff --git 
a/debian/patches/console-getty.service-don-t-start-when-dev-console-is-mis.patch
 
b/debian/patches/console-getty.service-don-t-start-when-dev-console-is-mis.patch
new file mode 100644
index 000..02eb1ca
--- /dev/null
+++ 
b/debian/patches/console-getty.service-don-t-start-when-dev-console-is-mis.patch
@@ -0,0 +1,49 @@
+From: Jan Pazdziora 
+Date: Fri, 13 Mar 2015 12:57:18 +0100
+Subject: console-getty.service: don't start when /dev/console is missing
+
+Create minimal image which runs systemd
+
+   FROM rhel7.1
+   RUN yum install -y /usr/bin/ps
+   ENV container docker
+   CMD [ "/usr/sbin/init" ]
+
+When you run the container without -t, the process
+
+   /sbin/agetty --noclear --keep-baud console 115200 38400 9600
+
+is not happy and checking the journal in the container, there is a stream of
+
+Mar 13 04:50:15 11bf07f59fff agetty[66]: /dev/console: No such file or 
directory
+Mar 13 04:50:25 11bf07f59fff systemd[1]: console-getty.service holdoff time 
over, scheduling restart.
+Mar 13 04:50:25 

Bug#832336: jessie-pu: package systemd/215-17+deb8u5

[Michael Biebl]

Am 24.07.2016 um 14:12 schrieb Michael Biebl:
>   * Fix prepare priority queue comparison function in sd-event.
> Otherwise a disabled event source can get swapped with an enabled one
> and cause a severe sd-event malfunction, breaking the event loop.
> (Closes: #789796)
> 
> http://anonscm.debian.org/cgit/pkg-systemd/systemd.git/commit/?h=jessie=3272c6b98
> 
> 
> This patch has been cherry-picked from upstream. It was part of v227, so
> has seen quite a bit of testing in unstable/stretch.
> It's a rather annoying/important bug fix, because once systemd got in
> this confused state, it was busy looping and could hog your CPU.

This is probably the most invasive change, so I've provided test
packages at [1]. The feedback via IRC on via email was positive [2]

Regards,
Michael


[1] https://people.debian.org/~biebl/systemd/jessie/
[2] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=789796#112
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=789796#117
-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?



signature.asc
Description: OpenPGP digital signature

Bug#832336: jessie-pu: package systemd/215-17+deb8u5

[Michael Biebl]

Package: release.debian.org
Severity: normal
Tags: jessie
User: release.debian@packages.debian.org
Usertags: pu

Dear release team,

for the upcoming jessie stable release, I'd like to update the systemd
package with the following fixes:

systemd (215-17+deb8u5) stable; urgency=medium

  * Use the right timeout for stop processes we fork. This ensures that
services are properly killed after a given timeout. (Closes: #813702)

http://anonscm.debian.org/cgit/pkg-systemd/systemd.git/commit/?h=jessie=036d0949c

This patch has been cherry-picked from upstream. It was part of v217, so
has seen quite a bit of testing in unstable/stretch. Not aware of any
regressions caused by that change.

  * Re-enable fsck -l. This avoids that systemd-fsck runs fsck in parallel for
file systems on the same disk. Parallel fsck can lead to suboptimal
behaviour on rotational hard disks. This requires util-linux 2.25, so bump
the dependency accordingly. (Closes: #824963)

http://anonscm.debian.org/cgit/pkg-systemd/systemd.git/commit/?h=jessie=aca6f1a0a

This patch has been cherry-picked from upstream. It was part of v217, so
has seen quite a bit of testing in unstable/stretch. Not aware of any
regressions caused by that change.


  * Don't reset log level to NOTICE if we get quiet on the kernel cmdline.
(Closes: #828006)

http://anonscm.debian.org/cgit/pkg-systemd/systemd.git/commit/?h=jessie=c1bf59723

This patch has been cherry-picked from upstream. It was part of v219, so
has seen quite a bit of testing in unstable/stretch. Not aware of any
regressions caused by that change.


  * Fix prepare priority queue comparison function in sd-event.
Otherwise a disabled event source can get swapped with an enabled one
and cause a severe sd-event malfunction, breaking the event loop.
(Closes: #789796)

http://anonscm.debian.org/cgit/pkg-systemd/systemd.git/commit/?h=jessie=3272c6b98


This patch has been cherry-picked from upstream. It was part of v227, so
has seen quite a bit of testing in unstable/stretch.
It's a rather annoying/important bug fix, because once systemd got in
this confused state, it was busy looping and could hog your CPU.

  * Update links to kernel.org cgroup documentation.
The systemd.resource-control man page had references to /cgroups/ which
moved to /cgroup-v1/. (Closes: #819970)

http://anonscm.debian.org/cgit/pkg-systemd/systemd.git/commit/?h=jessie=d5e94d3fb

This patch has been cherry-picked from upstream. It was part of v230.
Small documentation update with no regression potential.


  * Don't start console-getty.service when /dev/console is missing.
Avoids repeated unsuccessful start attempts of agetty inside (docker)
containers. (Closes: #829537)

http://anonscm.debian.org/cgit/pkg-systemd/systemd.git/commit/?h=jessie=51d6e161e

This patch has been cherry-picked from upstream. It was part of v220, so
has seen quite a bit of testing in unstable/stretch. Very low regression
potential.

  * getty-static.service: Only start if we have a working VC subsystem.
Use ConditionPathExists=/dev/tty0, the same check as in getty@.service,
to determine whether we have a functional VC subsystem and we should
start any gettys. (Closes: #824779)

http://anonscm.debian.org/cgit/pkg-systemd/systemd.git/commit/?h=jessie=95dc10016

getty-static.service is a Debian specific addition, so this change is
not an upstream cherry-pick. This fix hasn't been uploaded to unstable
yet, but is atm only committed to master and pending. So strictly
speaking it violates the stable upload policy.
The one-line fix seemed obvious enough though to include it in the
upcoming jessie release and not defer for another cycle.
If you have any concerns regarding this change, I'm happy to back it out
again. It's a small fix which mostly helps LXC containers, which don't
have a real VC subsystem but start some pseudo-ttys on tty1-tty4. So the
check for /dev/tty2 was wrong. We had a discussion with systemd upstream
who recommends to test for /dev/tty0 as indicator for a working VC
subsystem.

 -- Michael Biebl   Sun, 24 Jul 2016 11:58:28 +0200


The complete debdiff is attached as well.

Thanks for considering,
Michael

-- System Information:
Debian Release: stretch/sid
  APT prefers unstable-debug
  APT policy: (500, 'unstable-debug'), (500, 'unstable'), (200, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.6.0-1-amd64 (SMP w/4 CPU cores)
Locale: LANG=de_DE.UTF-8, LC_CTYPE=de_DE.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
diff --git a/debian/changelog b/debian/changelog
index 974bbb0..6568877 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,3 +1,30 @@
+systemd (215-17+deb8u5) stable; urgency=medium
+
+  * Use the right timeout for stop processes we fork. This ensures that
+services are properly killed after a given timeout. (Closes: #813702)
+  * Re-enable fsck -l. This