Unexpected registrar conflict on log rotation

[driskell]

Unexpected registrar conflict on log rotation

[driskell]

Thanks. Can you post the logs that occur before it started? Specifically, the ones regarding rotation.
I will look into this. Looks like a rename didn't get through to registrar.

[its-johnt]

I looked back a bit and apparently the issue started yesterday actually, here are the logs from the beginning of yesterday until the issue started. I notice it says that /var/log/syslog hasn't changed in 24 hours, but it's updated all the time every day, so maybe it still has the previous day's syslog file (which has since been rotated away) open.

Please let me know if you need anything else. Thanks for your help.

2014/08/26 00:02:19.033607 Launching harvester on new file: /var/log/postgresql/postgresql-2014-08-26.log
2014/08/26 00:02:19.036800 Started harvester at position 0 (requested 0): /var/log/postgresql/postgresql-2014-08-26.log
2014/08/26 06:27:06.553100 Stopping harvest of /var/log/upstart/log-courier.log; last change was 24h0m4s ago
2014/08/26 06:38:02.490053 Stopping harvest of /var/log/firewall.log; last change was 24h0m4s ago
2014/08/26 06:44:36.342593 Stopping harvest of /var/log/mysql.log; last change was 24h0m1s ago
2014/08/26 06:44:36.342653 Stopping harvest of /var/log/mysql/error.log; last change was 24h0m1s ago
2014/08/26 06:44:38.703017 Stopping harvest of /var/log/syslog; last change was 24h0m3s ago
2014/08/26 06:47:05.901974 Unexpected file truncation, seeking to beginning: /var/log/munin/munin-node.log
2014/08/26 06:47:10.699312 Launching harvester on rotated file: /var/log/syslog
2014/08/26 06:47:10.699552 Launching harvester on rotated file: /var/log/firewall.log
2014/08/26 06:47:10.699596 Launching harvester on rotated file: /var/log/mysql.log
2014/08/26 06:47:10.699698 Launching harvester on rotated file: /var/log/mysql/error.log
2014/08/26 06:47:10.699929 Launching harvester on rotated file: /var/log/upstart/log-courier.log
2014/08/26 06:47:10.703696 Started harvester at position 0 (requested 0): /var/log/syslog
2014/08/26 06:47:10.706723 Started harvester at position 0 (requested 0): /var/log/firewall.log
2014/08/26 06:47:10.706755 Started harvester at position 0 (requested 0): /var/log/mysql.log
2014/08/26 06:47:10.706780 Started harvester at position 0 (requested 0): /var/log/mysql/error.log
2014/08/26 06:47:10.706808 Started harvester at position 0 (requested 0): /var/log/upstart/log-courier.log
2014/08/26 06:47:10.706879 BUG: Unexpected registrar conflict detected for /var/log/syslog
2014/08/26 06:47:14.584157 BUG: Unexpected registrar conflict detected for /var/log/syslog
2014/08/26 06:47:19.584541 BUG: Unexpected registrar conflict detected for /var/log/syslog
2014/08/26 06:47:24.583812 BUG: Unexpected registrar conflict detected for /var/log/syslog

[driskell]

This was fixed a couple weeks ago it seems in my current work on monitoring features.

I hope to complete the monitoring stuff very soon so will get it pushed up in the next few days. If not, I'll back port the fix so it's in the main repository.

Apologies for the issue.

The impact is small - it just means if courier restarted in that 1 minute after rotation, it will potentially duplicate a tiny amount of logs that appeared in that minute.

[driskell]

Hi,

The fix for this should now be merged. If it happen again please reopen the issue but after some testing it should be resolved.

Sorry for the issue!

Jason