[erlang-questions] Fishing for best practices: distributed twin processes!
Tue Sep 25 20:05:39 CEST 2012
"Some time" is the net-tick-time, which defaults to 15 seconds I believe.
You can decrease that time by making it smaller (more frequent pings), at
the cost of (a potentially much greater) increase in network traffic.
The issue surrounds "half closed sockets" which is a general TCP problem
and less of an Erlang-specific issue. (eg, it is a problem all networked
applications have to deal with, including the upcoming Websockets stuff.)
On Tue, Sep 25, 2012 at 9:27 AM, Roberto Ostinelli <roberto@REDACTED>wrote:
> Thank you all for answers.
> The dying VM will of course not send anything, it's dead. Your node,
>> however, knows of all the links and monitors its processes has with other
>> nodes. It will detect that it has lost contact with the other node and and
>> then send the exit signals and monitor information to its local processes.
>> This may take some time before it happens.
> Robert, could you please clarify what 'some time' may mean? I'm wondering
> if it's better:
> - to have all processes linked between nodes, and receive the failover
> signal (may be thousands of them)
> - to have a single 'node monitoring' process which will then deal with
> a node failure
> Any input welcome :)
> erlang-questions mailing list
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the erlang-questions