#615: Cluster manager crash

Yep :frowning: Current workaround is to execute broctl start via cron every
minute. The manager and sometimes the proxy crash every hour or so
(but not on the hour).

... and the stack backtraces you get are still looking like the one
shown in this ticket? (I.e., it's not the 101 error in particular?)

My cluster is running stable now (with all processes running on
different boxes).


Ah, new error:

1317470292.374975 internal error in
/usr/local/bro-git/share/bro/base/protocols/http/./main.bro, line 269:
field value missing (HTTP::c$http$status_code)
/usr/local/bro-git/share/broctl/scripts/run-bro: line 60: 17933
Aborted (core dumped) nohup $mybro $@

Looks like another easy one involving the nested data structures not
checking for intermediate structs.

Yep, it was fixed yesterday in the repository.


Sorry, still not fixed as of yesterday:
internal error: unknown msg type 101 in Poll()
/usr/local/bro-git/share/broctl/scripts/run-bro: line 60: 6329
Aborted (core dumped) nohup $mybro $@