You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@storm.apache.org by Luke Forehand <lu...@networkedinsights.com> on 2014/06/04 18:10:08 UTC

tuples

Quick question, when a worker dies, while the supervisor is waiting for the timeout before re-launching the worker, what do upstream bolts do with the tuples they are attempting to send downstream?  Are they kept in a buffer until the downstream worker is alive again or do they get sent into the ether?

## downstream worker starting up
2014-06-04 11:05:09 b.s.d.supervisor [INFO] 49e93297-3082-4919-b47f-3426b362e448 still hasn't started
2014-06-04 11:05:10 b.s.d.supervisor [INFO] 49e93297-3082-4919-b47f-3426b362e448 still hasn't started

# upstream bolt attempting to reconnect to downstream
2014-06-04 11:04:22 b.s.m.n.Client [INFO] Reconnect ... [12]
2014-06-04 11:04:42 b.s.m.n.Client [INFO] Reconnect ... [13]
2014-06-04 11:05:02 b.s.m.n.Client [INFO] Reconnect ... [14]
2014-06-04 11:05:22 b.s.m.n.Client [INFO] Reconnect ... [15]
# re-established


Thanks,
Luke Forehand |  Networked Insights  |  Software Engineer