PR sent: https://github.com/mozilla-services/heka/pull/600
On Sun, Jan 5, 2014 at 5:45 PM, David Birdsong <[email protected]>wrote: > I answered my own question by reading the code. It appears there's no > reconnection logic on tcp_output.go. > > ...moving the conversation to github. > > > On Sun, Jan 5, 2014 at 2:35 PM, David Birdsong > <[email protected]>wrote: > >> What's the recommended way to restart a hekad aggregator? >> >> In trying out a hekad instance that forwards messages to another hekad >> instance, I see writer side throw errors when I restart the upstream side >> and never recover. >> >> Is there some way to trigger automatic reconnection? I'd settle for the >> writer exiting on a write failure so something like circusd could restart >> it. >> >> Basically this error on the writer side never resolves itself after the >> other instance fully restarts: >> 2014/01/05 22:39:00 Plugin 'aggregator_output' error: writing to >> svab35:5565: write tcp 10.0.1.35:5565: broken pipe >> >> >
_______________________________________________ Heka mailing list [email protected] https://mail.mozilla.org/listinfo/heka
