rsyslogd dumps core after adding aggregator relation
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
rsyslog (Juju Charms Collection) |
Expired
|
Undecided
|
Unassigned |
Bug Description
If I deploy the rsyslog charm to a lxc precise container, I can ssh in and see the rsyslogd process happily running.
If I then add an aggregator relation, rsyslogd is found to not be running.
Running rsyslogd manually, I find it dumping core.
root@stub-
rsyslogd: Could not open output pipe '/dev/xconsole' [try http://
Apr 2 19:40:14 stub-local-
Apr 2 19:40:14 stub-local-
Apr 2 19:40:14 stub-local-
Apr 2 19:40:14 stub-local-
Apr 2 19:40:14 stub-local-
Segmentation fault (core dumped)
Changed in rsyslog (Juju Charms Collection): | |
status: | New → Incomplete |
The trigger seems to be the final line of /etc/rsyslog. d/60-aggregator .conf:
$InputTCPServerRun 10514
Commenting out this line, rsyslogd will start up. However it won't be listening on port 10514, which I think is the port clients are supposed to be talking too (this just might be a documentation issue - the syslog interface has not been documented anywhere I can see).