You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Describe the bug
On a cluster of three nodes followed the Clustering and Upgrade documentation to upgrade the nodes.
Node A and Node C couldn't complete the postinst script of the packetfence package /var/lib/dpkg/info/packetfence.postinst .
Upgrade on Node B worked as expected.
To Reproduce
Upgrade a cluster from Packetfence 14.0 → 14.1 like described in the documentation. I don't know, yet, any circumstances that need to be met to trigger this bug.
Expected behavior
Upgrade should work on all of the nodes the same.
+ systemctl start packetfence-config
Job for packetfence-config.service failed because the control process exited with error code.
See "systemctl status packetfence-config.service" and "journalctl -xeu packetfence-config.service" for details.
Here's some debugging information:
root@packetfence-demo-c3:~# journalctl -S-30m | grep 'bind: address already in use' | tail -n1
Mär 03 12:22:37 packetfence-demo-c3 pfconfig-docker-wrapper[3711507]: docker: Error response from daemon: driver failed programming external connectivity on endpoint pfconfig (06e5b4e6275468f46ae3ca70bc3e2ac392c42950a553028e984c0ff9058186ab): Error starting userland proxy: listen tcp4 100.64.0.1:44444: bind: address already in use.
root@packetfence-demo-c3:~# netstat -pna | grep 44444
tcp 0 0 100.64.0.1:44444 100.64.0.1:44444 ESTABLISHED 3677237/pf-mariadb
root@packetfence-demo-c3:~# systemctl status packetfence-mariadb.service
● packetfence-mariadb.service - PacketFence MariaDB instance
Loaded: loaded (/lib/systemd/system/packetfence-mariadb.service; enabled; preset: enabled)
Active: active (running) since Mon 2025-03-03 11:59:05 CET; 29min ago
Main PID: 3677237 (pf-mariadb)
Status: "Taking your SQL requests now..."
Tasks: 11 (limit: 14298)
Memory: 298.9M
CPU: 3.701s
CGroup: /packetfence.slice/packetfence-base.slice/packetfence-mariadb.service
├─3677237 pf-mariadb
├─3677434 /bin/sh /usr/bin/mysqld_safe --defaults-file=/usr/local/pf/var/conf/mariadb.conf
├─3677611 /usr/sbin/mariadbd --defaults-file=/usr/local/pf/var/conf/mariadb.conf --basedir=/usr --datadir=/var/lib/mysql --plugin-dir=/usr/lib/mysql/plugin ->
└─3677612 logger -t mysqld -p daemon error
Mär 03 12:13:11 packetfence-demo-c3 packetfence[3677237]: pf-mariadb(3677237) ERROR: [1741000391.32968] Failed to connect to config service for namespace resource::cluste>
Mär 03 12:13:11 packetfence-demo-c3 pf-mariadb[3677237]: Can't connect to pfconfig on containers-gateway.internal:44444 : Connection refused
Mär 03 12:13:11 packetfence-demo-c3 packetfence[3677237]: pf-mariadb(3677237) ERROR: [1741000391.43328] Failed to connect to config service for namespace resource::cluste>
Mär 03 12:13:11 packetfence-demo-c3 pf-mariadb[3677237]: [1741000391.43328] Failed to connect to config service for namespace resource::cluster_hosts(DEFAULT), retrying
Mär 03 12:13:11 packetfence-demo-c3 pf-mariadb[3677237]: Can't connect to pfconfig on containers-gateway.internal:44444 : Connection refused
Mär 03 12:13:11 packetfence-demo-c3 pf-mariadb[3677237]: [1741000391.53675] Failed to connect to config service for namespace resource::cluster_hosts(DEFAULT), retrying
Mär 03 12:13:11 packetfence-demo-c3 packetfence[3677237]: pf-mariadb(3677237) ERROR: [1741000391.53675] Failed to connect to config service for namespace resource::cluste>
Mär 03 12:13:11 packetfence-demo-c3 pf-mariadb[3677237]: Can't connect to pfconfig on containers-gateway.internal:44444 : Connection refused
Mär 03 12:13:11 packetfence-demo-c3 pf-mariadb[3677237]: [1741000391.63774] Failed to connect to config service for namespace resource::cluster_hosts(DEFAULT), retrying
Mär 03 12:13:11 packetfence-demo-c3 packetfence[3677237]: pf-mariadb(3677237) ERROR: [1741000391.63774] Failed to connect to config service for namespace resource::cluste>
Trying to stop mariadb via systemctl didn't work on both nodes:
I didn't get any further trying to find out what is happening here.
To get on with the update I killed the broken mariadb process with pkill -9 pf-mariadb. In the other terminal with the stuck systemctl stop this resulted in Bus n/a: changing state RUNNING → CLOSED, the systemctl ended and the terminal ended in a prompt again.
Running apt-get install let the postinst script run again for the unconfigured packet and the configuration ended with
Installation complete
* Please fire up your Web browser and go to https://@ip_packetfence:1443 to complete your PacketFence configuration.
* Please stop your iptables service if you don't have access to configurator.
The text was updated successfully, but these errors were encountered:
Ok that I have seen it . I have not clue the why yet AND it is very hard to reproduce AND looks completely random.
The issue is that when packetfence-mariadb restarts, it takes the same port as the port of packetfence-config 44444.
The only (far from perfect) way I found until now it to check if the port is used by mariadb or not.
If I recall correctly, I killed the pid associated with the port 44444.
I will bring it back to the team.
Thanks
JeGoi
changed the title
upgrade 14.0 → 14.1 fails
packetfence-mariadb is using packetfence-config port (44444) when it restart during PacketFence upgrade (14.0 to 14.1)
Mar 4, 2025
JeGoi
changed the title
packetfence-mariadb is using packetfence-config port (44444) when it restart during PacketFence upgrade (14.0 to 14.1)
packetfence-mariadb is using packetfence-config port (44444) when it restart on PacketFence upgrade (14.0 to 14.1)
Mar 4, 2025
Describe the bug
On a cluster of three nodes followed the Clustering and Upgrade documentation to upgrade the nodes.
Node A and Node C couldn't complete the postinst script of the packetfence package /var/lib/dpkg/info/packetfence.postinst .
Upgrade on Node B worked as expected.
To Reproduce
Upgrade a cluster from Packetfence 14.0 → 14.1 like described in the documentation. I don't know, yet, any circumstances that need to be met to trigger this bug.
Expected behavior
Upgrade should work on all of the nodes the same.
Additional context
Running
dash -x /var/lib/dpkg/info/packetfence.postinst configure
fails with:Here's some debugging information:
Trying to stop mariadb via systemctl didn't work on both nodes:
The
systemctl
command hangs and doesn't proceed. Looking at the mariadb log in another terminal shows:The process doesn't seem to have ended and is still blocking tcp:44444 which pf-config is failing to bind to. Looking at the process I found:
I didn't get any further trying to find out what is happening here.
To get on with the update I killed the broken mariadb process with
pkill -9 pf-mariadb
. In the other terminal with the stucksystemctl stop
this resulted inBus n/a: changing state RUNNING → CLOSED
, thesystemctl
ended and the terminal ended in a prompt again.Running
apt-get install
let the postinst script run again for the unconfigured packet and the configuration ended withThe text was updated successfully, but these errors were encountered: