[Bro-Dev] [Bro-Commits] [git/bro] topic/robin/threading-fix: Reworking thread termination logic. (c3c6967)
seth at icir.org
Fri Jul 13 07:45:19 PDT 2012
On Jul 13, 2012, at 10:33 AM, Robin Sommer wrote:
> Without packet input, network time indeed follows wall clock (that's
> generally the case, i.e. the timestamps you see in the manager's own
> logs (i.e., not the ones received from workers) are also coming from
> network_time internally, but the I/O loop sets that to current_time()
> in that case).
Ah, of course. It's amazing how many scenarios that approach covers perfectly.
The only situation I can see trouble in the future is when I start bugging people about doing cluster based tracefile processing. ;) That will need a packet distributor that is keeping a steady heartbeat for the whole cluster anyway so it would still not cause a problem because the manager's clock would inherently need to be driven by that anyway.
International Computer Science Institute
(Bro) because everyone has a network
More information about the bro-dev