OK, this is as simple as the two MQ servers fighting over the same port.
root@n2:~# systemctl status garagemq.service
● garagemq.service - AMQP server in golang
Loaded: loaded (/usr/lib/systemd/system/garagemq.service; enabled; preset:
enabled)
Active: active (running) since Thu 2025-11-13 10:43:40 UTC; 3min 0s ago
Main PID: 1458 (garagemq)
Tasks: 13 (limit: 74274)
Memory: 14.8M (peak: 15.0M)
CPU: 3.103s
CGroup: /system.slice/garagemq.service
└─1458 /usr/bin/garagemq --config /etc/garagemq/config.yaml
Nov 13 10:43:40 n2 garagemq[1458]: time="2025-11-13T10:43:40Z" level=info
msg="Append exchange" name=amq.fanout type=fanout vhost=/
Nov 13 10:43:40 n2 garagemq[1458]: time="2025-11-13T10:43:40Z" level=info
msg="Append exchange" name=amq.topic type=topic vhost=/
Nov 13 10:43:40 n2 garagemq[1458]: time="2025-11-13T10:43:40Z" level=info
msg="Append exchange" name=amq.header type=headers vhost=/
Nov 13 10:43:40 n2 garagemq[1458]: time="2025-11-13T10:43:40Z" level=info
msg="Append exchange" name= type=direct vhost=/
Nov 13 10:43:40 n2 garagemq[1458]: time="2025-11-13T10:43:40Z" level=info
msg="Initialize exchanges..." vhost=/
Nov 13 10:43:40 n2 garagemq[1458]: time="2025-11-13T10:43:40Z" level=info
msg="Initialize queues..." vhost=/
Nov 13 10:43:40 n2 garagemq[1458]: time="2025-11-13T10:43:40Z" level=info
msg="Initialize bindings..." vhost=/
Nov 13 10:43:40 n2 garagemq[1458]: time="2025-11-13T10:43:40Z" level=info
msg="Load messages into queues" vhost=/
Nov 13 10:43:40 n2 garagemq[1458]: time="2025-11-13T10:43:40Z" level=info
msg="Server started" address="127.0.0.1:5672"
Nov 13 10:43:40 n2 systemd[1]: Started garagemq.service - AMQP server in golang.
Garagemq was started before and holds it.
In that state rabbitmq-server fails
root@n2:~# systemctl restart rabbitmq-server
Job for rabbitmq-server.service failed because the control process exited with
error code.
See "systemctl status rabbitmq-server.service" and "journalctl -xeu
rabbitmq-server.service" for details.
But stop the former and the latter works:
root@n2:~# systemctl stop garagemq
root@n2:~#
root@n2:~# systemctl restart rabbitmq-server
root@n2:~#
All fine now.
---
In that sence it is similar to "apt install nginx apache2" which both try to
grab port 80 and the latter will fail. But while usually a service should
always start with sane default, when they might conflict that is different.
In the apache/nginx example they are configured to be "tolerant" against a
failing start, compare:
/var/lib/dpkg/info/rabbitmq-server.postinst:
...
invoke-rc.d rabbitmq-server $_dh_action || exit 1
/var/lib/dpkg/info/apache2.postinst
...
deb-systemd-invoke $_dh_action 'apache2.service' >/dev/null || true
Should those two have the same behavior?
Maybe, but then the discussion as with the other cases is that it would hide
that one of the services didn't work on install.
garagemq already is tolerant to be the second:
deb-systemd-invoke $_dh_action 'garagemq.service' >/dev/null || true
** Tags added: server-triage-discuss
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2130487
Title:
package rabbitmq-server 3.12.1-1ubuntu1.2 failed to install/upgrade:
»installiertes post-installation-Skript des Paketes rabbitmq-
server«-Unterprozess gab den Fehlerwert 1 zurück
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/rabbitmq-server/+bug/2130487/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
