Bug#826044: Hangs in apt hook with a zombie - problem still exists in debian10

[zm5s-trnc]

On Wed, 11 May 2022 22:59:41 +0200 Thomas Liske  
wrote:

> Hi,
>
> you are using some ansible deployment? Could you share your ansible
> role?
>
> This is a long-standing bug and it feels like that affected users are
> using aptitude. I wonder if this is related - could you give it a try
> (force_apt_get task parameter)?

Yes, using Ansible.

There's nothing special in it.

- name: Debian | Ensure Debian upgraded
  apt:
    update_cache: yes
    upgrade: "yes"
  tags:
    - os_force_upgrade

Bug#826044: Hangs in apt hook with a zombie - problem still exists in debian10

[Thomas Liske]

Hi,

you are using some ansible deployment? Could you share your ansible
role?

This is a long-standing bug and it feels like that affected users are
using aptitude. I wonder if this is related - could you give it a try
(force_apt_get task parameter)?


Regards,
Thomas


On Tue, 2022-05-10 at 18:00 +0900, zm5s-trnc wrote:
> Ho great...
> 
> Found out that there's the same trouble on some debian9 hosts.
> I say some because not all servers did the blockage.
> 
> 
> # cat /etc/debian_version
> 9.13
> # dpkg -l | grep needrestart
> ii  needrestart 2.11-3+deb9u1 all  check 
> which daemons need to be restarted after library upgrades
> 
> 
>    ├─sshd,2094
>    │   ├─sshd,24499
>    │   │   └─sh,24594 -c /usr/bin/python 
> /root/.ansible/tmp/ansible-tmp-1652172426.49-4195-
> 68636559789086/AnsiballZ_apt.py 
> && sleep 0
>    │   │   └─python,24595 
> /root/.ansible/tmp/ansible-tmp-1652172426.49-4195-
> 68636559789086/AnsiballZ_apt.py
>    │   │   └─aptitude,24910 -y -o 
> Dpkg::Options::=--force-confdef -o Dpkg::Options::=--force-confold 
> safe-upgrade
>    │   │   ├─aptitude,30264 -y -o 
> Dpkg::Options::=--force-confdef -o Dpkg::Options::=--force-confold 
> safe-upgrade
>    │   │   │   └─sh,30265 -c test -x 
> /usr/lib/needrestart/apt-pinvoke && /usr/lib/needrestart/apt-pinvoke
> || true
>    │   │   │   └─needrestart,30266
> /usr/sbin/needrestart
>    │   │   │   └─(10-dpkg,30299)
>    │   │   └─{aptitude},24914
> 

Bug#826044: Hangs in apt hook with a zombie - problem still exists in debian10

[zm5s-trnc]

Ho great...

Found out that there's the same trouble on some debian9 hosts.
I say some because not all servers did the blockage.


# cat /etc/debian_version
9.13
# dpkg -l | grep needrestart
ii  needrestart 2.11-3+deb9u1 all  check 
which daemons need to be restarted after library upgrades



  ├─sshd,2094
  │   ├─sshd,24499
  │   │   └─sh,24594 -c /usr/bin/python 
/root/.ansible/tmp/ansible-tmp-1652172426.49-4195-68636559789086/AnsiballZ_apt.py 
&& sleep 0
  │   │   └─python,24595 
/root/.ansible/tmp/ansible-tmp-1652172426.49-4195-68636559789086/AnsiballZ_apt.py
  │   │   └─aptitude,24910 -y -o 
Dpkg::Options::=--force-confdef -o Dpkg::Options::=--force-confold 
safe-upgrade
  │   │   ├─aptitude,30264 -y -o 
Dpkg::Options::=--force-confdef -o Dpkg::Options::=--force-confold 
safe-upgrade
  │   │   │   └─sh,30265 -c test -x 
/usr/lib/needrestart/apt-pinvoke && /usr/lib/needrestart/apt-pinvoke || true

  │   │   │   └─needrestart,30266 /usr/sbin/needrestart
  │   │   │   └─(10-dpkg,30299)
  │   │   └─{aptitude},24914

Bug#826044: Hangs in apt hook with a zombie - problem still exists in debian10

[zm5s-trnc]

Poking around I found this, could there be a relationship ?

  ├─sshd,14441
  │   └─sh,14736 -c /usr/bin/python 
/root/.ansible/tmp/ansible-tmp-1650945060.61-15873-131215080055549/AnsiballZ_apt.py 
&& sleep 0
  │   └─python,14737 
/root/.ansible/tmp/ansible-tmp-1650945060.61-15873-131215080055549/AnsiballZ_apt.py
  │   └─aptitude,15251 -y -o Dpkg::Options::=--force-confdef -o 
Dpkg::Options::=--force-confold safe-upgrade
  │   ├─aptitude,21329 -y -o 
Dpkg::Options::=--force-confdef -o Dpkg::Options::=--force-confold 
safe-upgrade
  │   │   └─sh,21330 -c test -x 
/usr/lib/needrestart/apt-pinvoke && /usr/lib/needrestart/apt-pinvoke || true
  │   │   └─frontend,21331 -w 
/usr/share/debconf/frontend /usr/sbin/needrestart

  │   │   └─(needrestart,21335)
  │   └─{aptitude},15256


# /usr/sbin/needrestart
debconf: DbDriver "config": /var/cache/debconf/config.dat is locked by 
another process: Resource temporarily unavailable


# lsof /var/cache/debconf/config.dat
COMMAND    PID USER   FD   TYPE DEVICE SIZE/OFF   NODE NAME
frontend 21331 root    4uW  REG  254,2    31187 131278 
/var/cache/debconf/config.dat



A needrestart process is spawned but due to 
/var/cache/debconf/config.dat locked by frontend it is unable to process.

Of course may be unrelated.

Bug#826044: Hangs in apt hook with a zombie - problem still exists in debian10

[zm5s-trnc]

Tried to install debian11 needrestart-3.5 package, no problem regarding 
dependencies, but problem still the same.



  ├─sshd,13265
  │   └─sh,14854 -c /usr/bin/python 
/root/.ansible/tmp/ansible-tmp-1650508029.18-25277-93685447926498/AnsiballZ_apt.py 
&& sleep 0
  │   └─python,14855 
/root/.ansible/tmp/ansible-tmp-1650508029.18-25277-93685447926498/AnsiballZ_apt.py
  │   └─aptitude,15365 -y -o Dpkg::Options::=--force-confdef -o 
Dpkg::Options::=--force-confold safe-upgrade
  │   ├─aptitude,21577 -y -o 
Dpkg::Options::=--force-confdef -o Dpkg::Options::=--force-confold 
safe-upgrade
  │   │   └─sh,21578 -c test -x 
/usr/lib/needrestart/apt-pinvoke && /usr/lib/needrestart/apt-pinvoke || true
  │   │   └─frontend,21579 -w 
/usr/share/debconf/frontend /usr/sbin/needrestart

  │   │   └─(needrestart,21583)
  │   └─{aptitude},15381

needrestart,21583 gets zombie-fied

We can move on if we kill the parent process frontend,21579