Hi all,
This problem is similar to one already discussed before, I know, but as no
solution was given, I try to put my case with as many details as I just
can provide, and kindly ask the gurus to help me.
Environment: Debian 2.0, kernel 2.0.36 (ie. I use kernel ax25),
ax25-utils-2.1.42a.glibc, XFBB 7.00g tied direct to kernel ax25.
The pain is, that sometimes a link got stuck and though it can't be seen
from the application side (XFBB), it shows up in netstat --ax25. This
prevents that station to connect again untill I reboot the whole system.
No axctl kill helps, apparently no way to get rid of it but only
restarting everything. The state of such a link seems always to be
"LISTENING", while Recv-Q is -196.
If I stop solely XFBB, then device changes to "???" (three times question
mark) in netstat response, but the rest of the line is the same. If I
restart XFBB, then netstat keeps on showing "???" in device column for the
stuck station.
Below you find one such case, when HG8GL-0 is stuck here.
So, any hint on what gives this? And what to do against? Or to get rid of
such, except the reboot?
Thanks for your help.
#netstat --ax25
Active AX.25 sockets
Dest Source Device State Vr/Vs Send-Q Recv-Q
HA3PG-0 HA5OB-0 ax3 RECOVERY 002/007 3780 0
* HA5OB-0 ax3 LISTENING 000/000 0 0
* HA5OB-0 scc3 LISTENING 000/000 0 0
* HA5OB-0 ax1 LISTENING 000/000 0 0
HG8GL-0 HA5OB-0 ax1 LISTENING 000/000 0 -196
#cat /proc/net/ax25
dest_addr src_addr dev st vs vr va t1 t2 t3 idle n2 rtt wnd
paclen Snd-Q Rcv-Q Inode
HA3PG-0 HA5OB-0 ax3 4 7 2 5 4/006 0/02 0/300 16/016 1/10 3 2
230 3780 0 146611
* HA5OB-0 ax3 0 0 0 0 0/006 0/02 0/300 0/016 0/10 3 2
256 0 0 86074
* HA5OB-0 scc3 0 0 0 0 0/005 0/01 0/300 0/016 0/10 2 7
256 0 0 86073
* HA5OB-0 ax1 0 0 0 0 0/005 0/01 0/300 0/016 0/10 2 7
256 0 0 86071
HG8GL-0 HA5OB-0 ax1 0 0 0 0 0/005 0/01 0/300 0/016 0/10 2 7
256 0 -196 0
#axctl bbs hg8gl ha5ob -kill
axctl: SIOAX25CTLCON: Transport endpoint is not connected
(Note: "bbs" is the name of "ax1" device as per /etc/ax25/axports.)
73 , Zoli
PS. Maybe I should drill a few holes big enough to make those stuck data
flow out of the system? ;-)
--------------------------------------------------------------
Zoltan (Zoli) Meszaros AX25 : HA5OB @ HA5OB.HUN.EURO
Budapest, Hunyadi ter 10. ampr-net : [EMAIL PROTECTED]
H-1067 email : [EMAIL PROTECTED]
HUNGARY WEB : http://www.qsl.net/ha5ob
