-----Message d'origine-----
De : Stefan Gybas <[EMAIL PROTECTED]>
@ : [EMAIL PROTECTED] <[EMAIL PROTECTED]>
Date : mardi 22 janvier 2002 21:19
Objet : Re: iuvc, debian, 2.4.16 ... Problem !
>Laurent SIKSOUS wrote:
>
>> "kernel: iucv_add_pathid: Pathid being used, error."
>
>
>Are you trying to establish multiple connections from one guest to the same
>peer?
No i don't. but from one guest to multiple peers wich in turn connect to
others.
>
>Please post the options you are passing to the netiucv module (from
>/etc/modules.conf or the command line) and the contents of
>/etc/network/interfaces, otherwise nobody will probably be able to help
you.
You're right. I must give more precise information. I am using an MP3000 who
runs VM 2.4.
In /etc/modules.conf, i have:
options iucv=TCPIP:LINUX1:LINUX2...
my /etc/network/interfaces looks like this:
# TCPIP
iface iucv0 inet static
address x.x.x.14
netmask 255.255.255.255
pointopoint x.x.x.13
# Interface iucv pour LINUX1
iface iucv1 inet static
address x.x.x.14
netmask 255.255.255.255
pointopoint x.x.x.1
# Interface iucv pour LINUX2
iface iucv1 inet static
address x.x.x.14
netmask 255.255.255.255
pointopoint x.x.x.2
and so on...
>
>You can also try to bring up the interfaces step by step, i.e. remove the
>auto statements from /etc/network/interfaces and enter "ifup iucv0",
>"ifup iucv1", ... until the error occurs. Then check the configuration of
>this interface in /etc/network/interfaces and /proc/net/iucv/iucv<X>/*.
(doing the job manually is the only way for me to make the guests finally
talking to each other)
Here's what I get in /proc/net/iucv/iucv<x>/statistics:
(after reboot with no auto satement for iucv interfaces)
Device FSM state: Stopped
Connection FSM state: Stopped
Max. TX buffer used: 0
Max. chained SKBs: 0
TX single write ops: 0
TX multi write ops: 0
Netto bytes written: 0
Max. TX IO-time: 0
(then after: ifup iucv1)
Device FSM state: StartWait
Connection FSM state: SetupWait
Max. TX buffer used: 0
Max. chained SKBs: 0
TX single write ops: 0
TX multi write ops: 0
Netto bytes written: 0
Max. TX IO-time: 0
(But at the other side, i get)
Jan 23 16:32:06 LINUX1 kernel: Device iucv0: Connection to user LINUX2 is up
Jan 23 16:32:06 LINUX1 kernel: iucv0: Connection to user LINUX2 is down
or,
kernel: iucv_add_pathid: Pathid being used, error.
and occasionally,
rc=-22
At last, I can activate the connections (but sometimes I even have to reload
the netiucv module) and , like i said before, i should be able to do it
without the silly effort of opening a session at the two sides and play with
commands.
I have an iucv subnet of 10 SUSE 7.0 virtual machines connected together
with kernel 2.2.16 and working very well. Until I find out what's happening,
is it reasonnable, would it be a good solution to compile such a kernel (for
my production needs) and keep on using Debian/testing as the rest of the
system ?
Thanks for your answer,
Laurent.
>
>--
>Stefan Gybas
>
>
