On Fri, Nov 05, 1999 at 09:49:30AM -0700, Bob Nielsen wrote: > > Since the latest potato update to diald 0.99.1-0.1, I have found that > after the link goes down, I have to kill diald and restart it before it > will dial again. /var/log/ppp.log shows: > > Nov 5 07:05:36 nielsen diald[9081]: Closing down idle link. > Nov 5 07:05:36 nielsen diald[9081]: start sl0: SIOCSIFMETRIC: Invalid > argument > Nov 5 07:05:36 nielsen diald[9081]: bind snoopfd: Bad file descriptor > Nov 5 07:05:36 nielsen pppd[9095]: Terminating on signal 2. > Nov 5 07:05:54 nielsen pppd[9095]: Hangup (SIGHUP) > Nov 5 07:05:54 nielsen pppd[9095]: Modem hangup > Nov 5 07:05:54 nielsen pppd[9095]: Connection terminated. > Nov 5 07:05:54 nielsen pppd[9095]: Connect time 14.2 minutes. > Nov 5 07:05:54 nielsen pppd[9095]: Sent 0 bytes, received 0 bytes. > Nov 5 07:05:54 nielsen pppd[9095]: Exit. > Nov 5 07:05:54 nielsen diald[9081]: start sl0: SIOCSIFMETRIC: Invalid > argument > Nov 5 07:05:54 nielsen diald[9081]: Disconnected. Call duration 865 seconds. > Nov 5 07:05:54 nielsen diald[9081]: IP transmitted 18201 bytes and received > 239701 bytes. > Nov 5 07:05:55 nielsen diald[9081]: Delaying 5 seconds before clear to dial. > > As far as I can tell, this is similar to what I was seeing with the > earlier version, except for the SIOCSIFMETRIC error, which also occurs > when the connection is made (previously there was a SIOCDELRT error > instead). > > I also get a message "unknown option 'reroute'" when the connection is > initiated. I don't see anything in the changelog about reroute being > discontinued. > > Prior to the latest upgrade, diald ran flawlessly for me. Running > > /etc/init.d/diald stop ; /etc/init.d/diald start > > will reset things so diald works again, but this shouldn't be > necessary. Does anyone know what changed here and what to do about it? > > System is potato (nearly up-to-date, except for a few packages which > won't update correctly, like mutt and console-data), kernel 2.0.38. >
I noticed the same thing with a 2.0.34 kernel. Booting a 2.2.10 kernel seems to have cured things. HTH Your Pal Dave -- Dave Thayer Denver, Colorado USA [EMAIL PROTECTED]
