We rencelty updated our docker image to use a more recent version of pushpin. Previously we had Condure + Zurl, and we switched to connmgr with the new puhspin version. That means that since then we have a new connmgr.log file, we missed that and had permissions issues. We saw that connmgr was crashing on startup :
pushpin@xxxxxx:/$ pushpin --verbose
[DEBUG] 2026-07-29 15:00:42.234 starting...
[INFO] 2026-07-29 15:00:42.234 using config: /etc/pushpin/pushpin.conf
[INFO] 2026-07-29 15:00:42.243 starting connmgr
[DEBUG] 2026-07-29 15:00:42.243 running: pushpin-connmgr --log-level=3 --buffer-size=8192 --stream-maxconn=50000 --listen=0.0.0.0:7999,stream --zclient-stream=ipc:///var/run/pushpin/pushpin-connmgr --zserver-stream=ipc:///var/run/pushpin/pushpin-connmgr-client --deny-out-internal
[ERR] 2026-07-29 15:00:42.243 connmgr: Error running: pushpin-connmgr
[INFO] 2026-07-29 15:00:42.243 stopping proxy
[INFO] 2026-07-29 15:00:42.243 stopping handler
[INFO] 2026-07-29 15:00:42.243 starting proxy
[DEBUG] 2026-07-29 15:00:42.243 running: pushpin-proxy --config=/etc/pushpin/pushpin.conf --ipc-prefix=pushpin- --logfile=/var/tracim/logs/pushpin/pushpin-proxy.log --loglevel=3
[INFO] 2026-07-29 15:00:42.245 starting handler
[DEBUG] 2026-07-29 15:00:42.245 running: pushpin-handler --config=/etc/pushpin/pushpin.conf --ipc-prefix=pushpin- --logfile=/var/tracim/logs/pushpin/pushpin-handler.log --loglevel=3
[DEBUG] 2026-07-29 15:00:42.247 started
Unfortunately no log message is saying the crash cause.
In our configuration it was indeed caused by connmgr.log not writable by the pushpin user.
Can a console log be added when starting pushpin to indicate log file isn't writable ?
We rencelty updated our docker image to use a more recent version of pushpin. Previously we had Condure + Zurl, and we switched to connmgr with the new puhspin version. That means that since then we have a new connmgr.log file, we missed that and had permissions issues. We saw that connmgr was crashing on startup :
Unfortunately no log message is saying the crash cause.
In our configuration it was indeed caused by connmgr.log not writable by the pushpin user.
Can a console log be added when starting pushpin to indicate log file isn't writable ?