From: Bob Purdon <bobp@dont-contact.us>
Date: Fri, 5 Jun 1998 16:44:55 +1000 (EST)

Hi All,

Has anyone else noticed a dramatic increase in the number of TIMEOUT_*
entries in their logs with 1.2b21 and 1.2b22?

In our case it's using 1.2b20 as a sibling, although I've seen the same
behaviour using 1.2b20 as a parent also.

A typical log line is as follows. Note that our neighbour timeout is set
to 2 seconds. The neighbour concerned is 3ms away and not heavily loaded.

897027367.571 724 TCP_MISS/200 1000 GET -

It seems to be better when I force it to talk to a 1.2b22 parent, but
there's still quite a few timeouts. With 1.2b20, configured identically,
we saw very few.

Am I missing something here?


Bob Purdon,
Technical Manager,
Southern Internet Services.

PGP key at: http://www.southcom.com.au/~bobp/bobp.asc
Received on Thu Jun 04 1998 - 22:46:28 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:40:37 MST