pserv "connection refused" errors spamming its log file

[Julian Edwards]

I installed upgraded packages and after the installation finished, pserv was reporting:

2014-10-10 15:13:22+1000 [ClusterClient,client] Event-loop 'maas:pid=23213' authenticated.
2014-10-10 15:13:22+1000 [ClusterClient,client] Event-loop 'maas:pid=23213' authenticated.
2014-10-10 15:13:22+1000 [ClusterClient,client] Event-loop 'maas:pid=23212' authenticated.
2014-10-10 15:13:22+1000 [ClusterClient,client] Event-loop 'maas:pid=23212' authenticated.

Followed by:

2014-10-10 15:13:23+1000 [Uninitialized] Event-loop maas:pid=22266 (127.0.0.1:42097): Connection was refused by other side: 111: Connection refused.
2014-10-10 15:13:23+1000 [Uninitialized] Event-loop maas:pid=22266 (127.0.0.1:42097): Connection was refused by other side: 111: Connection refused.
2014-10-10 15:13:23+1000 [Uninitialized] Event-loop maas:pid=22266 (10.0.0.9:42097): Connection was refused by other side: 111: Connection refused.
2014-10-10 15:13:23+1000 [Uninitialized] Event-loop maas:pid=22266 (10.0.0.9:42097): Connection was refused by other side: 111: Connection refused.
2014-10-10 15:13:23+1000 [Uninitialized] Event-loop maas:pid=22266 (192.168.122.1:42097): Connection was refused by other side: 111: Connection refused.
2014-10-10 15:13:23+1000 [Uninitialized] Event-loop maas:pid=22266 (192.168.122.1:42097): Connection was refused by other side: 111: Connection refused.

The latter opes here repeating over and over.

Note that 10.0.0.9 is the actual IP address of my maas host. The UI reports that the cluster is connected OK.

[Julian Edwards (julian-edwards)]

Targeting to next since everything appears to actually work, but the log noise is a nuisance. Any ideas Gavin?

[Gavin Panella (allenap)]

Did it shut up after about 5 minutes? If so, it's probably because the event-loops were not able to stop cleanly and remove their host:port entries from the database. This is bug 1350786.

[Julian Edwards (julian-edwards)]

On Friday 10 Oct 2014 09:21:15 you wrote:
> Did it shut up after about 5 minutes? If so, it's probably because the
> event-loops were not able to stop cleanly and remove their host:port
> entries from the database. This is bug 1350786.

Oh I didn't check. I'll look again next time I run up my rig, please feel
free to dupe this if you're confident, I can easily undupe if necessary.

[Andreas Hasenack (ahasenack)]

Also getting tons of these connection refused messages in the logs.

[Julian Edwards (julian-edwards)]

Un-duping from bug 1350786 as this seems to be a result of it trying to connect to 127.0.0.1 (and the link-local address, from memory).