• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

Issue Plesk Update 18.0.39 fails when Bind not installed

TimReeves

Regular Pleskian
Has happened already at 2 servers. In general I don't install Bind, don't need it. Plesk Updates have always worked fine. Now suddenly they don't, same at both servers. From the Updater log:

Reading state information...
bind9 is already the newest version (1:9.11.5.P4+dfsg-5.1+deb10u6).

That's strange - bind9 should not be installed (Option is off in Plesk)

Setting up bind9 (1:9.11.5.P4+dfsg-5.1+deb10u6) ...
ls: cannot access '/etc/bind/rndc.key': No such file or directory
bind9-pkcs11.service is a disabled or a static unit not running, not starting it.
bind9-resolvconf.service is a disabled or a static unit not running, not starting it.
Job for bind9.service failed because the control process exited with error code.
See "systemctl status bind9.service" and "journalctl -xe" for details.
invoke-rc.d: initscript bind9, action "restart" failed.
* bind9.service - BIND Domain Name Server
Loaded: loaded (/lib/systemd/system/bind9.service; disabled; vendor preset: enabled)
Active: failed (Result: exit-code) since Sat 2021-10-30 03:03:52 CEST; 10ms ago
Docs: man:named(8)
Process: 26637 ExecStart=/usr/sbin/named $OPTIONS (code=exited, status=1/FAILURE)

dpkg: error processing package bind9 (--configure):
installed bind9 package post-installation script subprocess returned error exit status 1

Errors were encountered while processing:
bind9
E: Sub-process /usr/bin/dpkg returned an error code (1)

ERROR: An error occurred on attempt to install packages.
Attention! Your software might be inoperable.
Visit https://support.plesk.com/ to search for a solution.
 
Hello!
I can not reproduce it on Debian 10 (update Plesk 18.0.38 -> 18.0.39 w/o BIND)
Could you provide more information?
cat /etc/apt/sources.list
 
Thanks @yborzykina ,
here we go:

# deb http://asi-fs-d.contabo.net/debian buster main

deb http://asi-fs-d.contabo.net/debian buster main non-free contrib
deb-src http://asi-fs-d.contabo.net/debian buster main non-free contrib

deb Index of /debian-security buster/updates main contrib non-free
deb-src Index of /debian-security buster/updates main contrib non-free

# buster-updates, previously known as 'volatile'
deb http://asi-fs-d.contabo.net/debian buster-updates main contrib non-free
deb-src http://asi-fs-d.contabo.net/debian buster-updates main contrib non-free
## These are official Contabo mirrors providing faster updates for your server ##
## (the files from the official servers are synced to it) ##

What I notice is that the Provider Contabo uses its own mirrors. And the problem I have only occurs with Contabo servers and Debian 10. I have a Deb 9 server at Contabo which updated fine, and a Debian 10 server at another provider which also updated fine.

I just tried to update one the problem servers again, the problem remains.

What do you think?
 
Hello!
please run the command: rndc-confgen -r /dev/urandom -a and then try to upgrade with --debug
Let me check autoinstaller.log.
 
Hello,

I have the same issue, trying to update to 18.0.39 Update 2 from 18.0.39 Update 1.

Relevant part of the log file:

Code:
Setting up bind9 (1:9.11.5.P4+dfsg-5.1+deb10u6) ...
bind9-pkcs11.service is a disabled or a static unit not running, not starting it.
bind9-resolvconf.service is a disabled or a static unit not running, not starting it.
Job for bind9.service failed because the control process exited with error code.
See "systemctl status bind9.service" and "journalctl -xe" for details.
invoke-rc.d: initscript bind9, action "restart" failed.
* bind9.service - BIND Domain Name Server
   Loaded: loaded (/lib/systemd/system/bind9.service; disabled; vendor preset: enabled)
   Active: failed (Result: exit-code) since Fri 2021-11-19 16:24:26 CET; 7ms ago
     Docs: man:named(8)
  Process: 7957 ExecStart=/usr/sbin/named $OPTIONS (code=exited, status=1/FAILURE)

Nov 19 16:24:26 einstein.williamblondel.fr named[7958]: found 4 CPUs, using 2 worker threads
Nov 19 16:24:26 einstein.williamblondel.fr named[7958]: using 2 UDP listeners per interface
Nov 19 16:24:26 einstein.williamblondel.fr named[7958]: using up to 4096 sockets
Nov 19 16:24:26 einstein.williamblondel.fr named[7958]: loading configuration from '/etc/named.conf'
Nov 19 16:24:26 einstein.williamblondel.fr named[7958]: open: /etc/named.conf: permission denied
Nov 19 16:24:26 einstein.williamblondel.fr named[7958]: loading configuration: permission denied
Nov 19 16:24:26 einstein.williamblondel.fr named[7958]: exiting (due to fatal error)
Nov 19 16:24:26 einstein.williamblondel.fr systemd[1]: bind9.service: Control process exited, code=exited, status=1/FAILURE
Nov 19 16:24:26 einstein.williamblondel.fr systemd[1]: bind9.service: Failed with result 'exit-code'.
Nov 19 16:24:26 einstein.williamblondel.fr systemd[1]: Failed to start BIND Domain Name Server.
dpkg: error processing package bind9 (--configure):
 installed bind9 package post-installation script subprocess returned error exit status 1

I'm not sure what to do.
Thanks!
 
hello @wblondel ,

how did you executed plesk upgrade ?

was it through Plesk UI or through command line ?

or it is automatic upgrade by Plesk itself and you have got result only, when it was finished ?
 
also, could you please provide result of execution
dpkg -S /etc/named.conf

@TimReeves ,
did you have bind installed but disabled before upgrade? or it was removed at all after Plesk installation ?

if the second: how got you removed it ?
 
I found that bind was installed on the two servers which could not update. In Plesk the option was deselected, but the packages were installed nevertheless. I deleted the packages and then the update worked.
Perhaps Plesk could check if bind9 and bind9utils are installed and work round the issue, or issue a notification to admin?
Thanks for your support.
 
hi again.

I was not able to reproduce such error in case legitimate situation:

- plesk was initially installed without bind component DNS management was disabled in plesk and Bind-related packages were removed(bind configuration also was removed)
- package plesk-dns-bind-driver was removed (bind configuration also was removed)
- package bind9 was removed(with bind configuration files/dirictories) which leads to removing package plesk-dns-bind-driver

in all cases above Plesk does not try to manipulate with Bind service.

provided error
open: /etc/named.conf: permission denied

mean that bind9.service was broken but packages were not correctly removed from the system.
Additionally this error may mean that dns management has not been fully disabled in Plesk (otherwise Plesk will not try to configure it, which leads to service (re)start )

Most probably service was disabled by systemctl and plesk was ignoring these problems.
But during upgrade it also tried to update packages related to bind and service failed to start due to broken configuration.

I was able to reproduce such situation by
1. installing outdated Bind packages (previous versions)
2. disabling bind.service
3. removing /etc/bind directory
4. start upgrade plesk

after actions above Plesk upgrade failed in my case (similar error with inability to start bind.service).

We'll try to figure out together with Developers how we can catch such errors, but for now I can't provide any ETA.
All that we can recommend is to remove componet bind with plesk installer
or
try to repair bind configuration by invoking
rndc-confgen -r /dev/urandom -a
and
plesk repair dns
 
Back
Top