Q: where do these tcp connections come from ?
Reto Kramer
kramer@REDACTED
Mon Sep 17 08:32:53 CEST 2001
I'm surprised to find a number of tcp connections that I cannot quite figure
out why they were created. Maybe someone familiar with the inner workings of
the net_kernel can help me out?
Two 5.0.2 nodes are running on the same Win2K box. I can see the 4 tcp links
from the two nodes to the port daemon (1/2 and 3/4 in the netstat below).
Afterward I ping/pong node b from node a and see two more connections (5/6).
All is fine. Any further ping/pong among a and b is multiplex over the
existing connections.
Now the funny thing is that for each "world_list" of the form:
net_adm:world_list([deepthought],verbose).
Pinging b@REDACTED -> pong
Pinging a@REDACTED -> pong
[b@REDACTED,a@REDACTED]
I find a new connection that was opened and waits to be cleaned up (3 times
world_list and I got 7,8 and 9). Please note that no new machine and no new
node was involved - it's all on the deepthought box. The connections look as
if they were made to the port daemon (4306).
Q: is there a specific reason why the existing connection (e.g. 3) could not
have been used for the world_list traffic?
Q: from a naive look at world_list, ping and gen:call I cannot figure out
why world_list behaves differently from a plain "ping" in terms of
connection multiplexing. Can someone explain why?
$ netstat
Proto Local Address Foreign Address State
1 TCP deepthought:1495 deepthought:4369 ESTABLISHED
2 TCP deepthought:4369 deepthought:1495 ESTABLISHED
3 TCP deepthought:1501 deepthought:4369 ESTABLISHED
4 TCP deepthought:4369 deepthought:1501 ESTABLISHED
5 TCP deepthought:1493 deepthought:1505 ESTABLISHED
6 TCP deepthought:1505 deepthought:1493 ESTABLISHED
7 TCP deepthought:4369 deepthought:1504 TIME_WAIT
8 TCP deepthought:4369 deepthought:1507 TIME_WAIT
9 TCP deepthought:4369 deepthought:1508 TIME_WAIT
cheers,
- Reto
More information about the erlang-questions
mailing list