hey,
For this server the script does use the generic kernel and i believe the namechange is due to the
-b parameter in the config command
config -b ${HOSTNAME} ...
anyway I am using the default kernel.
The problem however remains with rsh/rcp where concurrent transmissions acuse some of them to either lose connection or simply fail with 'no route to host' while there are no connection problems at all as far as i can see.
I did find something through google about tcp connections being kept open and thus still holding the port for a while
and that it could give problems when another process tries to use that port. Is it possible that rsh/rcp always tries to
use the same source port on the host and that new connections fail? Seems unlikely to me since the error is 'no route
to host' but still...
Anyway who can shed some light on this? Thanks, Stefan
Stuart Henderson wrote:
Hi Stefan,
--On 02 May 2005 17:19 +0200, Stefan Castille wrote:
Is there any more information i need to provide to solve this issue, any setting i can change that will solve this problem?
OpenBSD 3.6-stable (magnesium) #0: Fri Feb 25 12:22:40 CET 2005 [EMAIL PROTECTED]:/usr/obj/magnesium
First, try a generic kernel.
It is a generic kernel, I only have a script that automatically checks CVS for patches, downloads them and then does a full recompile, there is nothing custom about the machine at all.
It can't be exactly the generic kernel, since it says 'magnesium' after the version number. If nothing else has changed (i.e. 'diff GENERIC magnesium' shows the change of name, and nothing else) then maybe follow-up to the list and mention that.
<http://www.openbsd.org/faq/faq5.html#Why>
