Second, that means your real problem is with the inability to restart. I don't use Bering myself, but I would expect it to include a mechanism for restoring dropped connections, either automatically or on demand. You might check the docs a bit more closely on that score.
Third, your immediate problem is that when the connection goes down, your keepalive script does not restart it. Nor does your executing the same command (/etc/init.d/networking restart) from the command line. Since that is the command to stop, then restart, all interfaces, this command should work ... or at least it should do *something*.
So ... please report this part in more detail than "but it didn't restart until the modem was powercycled". Use the SR FAQ to learn the set of diagnostics we want to see (in this case, run before and after you execute the command).
Also please clarify what you mean by "eventually the system stops trying".
The problem may be in the details of how your PPPoE connection gets set up, in particular how the rotuer communicates locally with the DSL modem. PPPoE setup requirements imposed by ISPs vary a lot (who is your ISP, BTW? I know PacBell uses the Enternet 300 Windows PPPoE client, or did bak when I struggled with a PPPoE connection), and dealing with your problem may require information on the details of your ISP's setup instructions (for Windows).
At 10:08 AM 1/26/03 -0500, Arcana wrote:
Hello,
I am having problems with my ISP disconnecting me from the web.
My ISP is a DSL provider that uses PPPoE. The CD that they provided me for a
Windows installation is the Enternet 300 client.
I use a Bering 1.0 stable distribution.
ppp0 is the PPP interface, and it is connected to eth0.
eth1 leads to my internal network.
Initially, ppp0 was disconnecting every night after a certain period of
inactivity (something along the lines of 2 hours). I fixed this by adding a
script in my cron that would ping a host 10 times every 15 minutes. After I
did this, the disconnects stopped. (If anyone has a better solution for
keeping DSL alive please let me know). When I would renew, I would get the
same IP address (probably because lease time was not expired).
The other day, I woke up and the Internet did not work. I tried numerous
times to restart the connection by running '/etc/init.d/networking restart'
but it didn't restart until the modem was powercycled. Once I recovered the
connection the IP changed, leading me to suspect that when my IP changes the
PPPoE doesn't pick it up (or similar) and ends up disconnecting.
One person I spoke with who has a decent router/firewall setup recommended
that I try Roaring Penguin PPPoE (he had no problems with his DSL
disconnecting). I've looked a little at the PPPoE setup in Bering, and I
read that using the pppoe executable to help route packets over ppp was
supposed to be a stopgap measure until PPPoE was supported in the kernel
modules (or something).
Here's the log surrounding the incident... maybe someone can get a little more
information out of it than I can. I've read over the DSL/PPPoE notes and the
PPP notes on LEAF and am currently reviewing the DSL-HOWTO for Linux to see
if I can get more ideas.
Jan 24 07:50:01 firewall /USR/SBIN/CRON[10713]: (root) CMD
(/etc/keepalive_script)
Jan 24 07:59:25 firewall pppd[9920]: LCP terminated by peer
Jan 24 07:59:25 firewall pppd[9920]: Couldn't increase MTU to 1500.
Jan 24 07:59:25 firewall pppd[9920]: Couldn't increase MRU to 1500
Jan 24 07:59:25 firewall pppd[9920]: Modem hangup
Jan 24 07:59:25 firewall pppd[9920]: Connection terminated.
Jan 24 07:59:25 firewall pppd[9920]: Connect time 10068.8 minutes.
Jan 24 07:59:25 firewall pppd[9920]: Sent 50288437 bytes, received 209080228
bytes.
Jan 24 07:59:25 firewall pppd[9920]: Doing disconnect
Jan 24 07:59:55 firewall pppd[9920]: Sending PADI
Jan 24 07:59:55 firewall pppd[9920]: HOST_UNIQ successful match
Jan 24 07:59:55 firewall pppd[9920]: HOST_UNIQ successful match
Jan 24 07:59:55 firewall pppd[9920]: Got connection: 1b7f
Jan 24 07:59:55 firewall pppd[9920]: Connecting PPPoE socket:
00:90:1a:40:44:2c 7f1b eth0 0x807c2c8
Jan 24 07:59:55 firewall pppd[9920]: using channel 21
Jan 24 07:59:55 firewall pppd[9920]: Using interface ppp0
Jan 24 07:59:55 firewall pppd[9920]: Connect: ppp0 <--> eth0
Jan 24 07:59:55 firewall pppd[9920]: Couldn't increase MTU to 1500.
Jan 24 07:59:55 firewall pppd[9920]: Couldn't increase MRU to 1500
Jan 24 07:59:56 firewall pppd[9920]: Couldn't increase MTU to 1500.
Jan 24 07:59:56 firewall pppd[9920]: Couldn't increase MRU to 1500
Jan 24 07:59:56 firewall pppd[9920]: Couldn't increase MTU to 1500.
Jan 24 07:59:56 firewall pppd[9920]: Couldn't increase MRU to 1500
Jan 24 07:59:56 firewall pppd[9920]: Cannot determine ethernet address for
proxy ARP
Jan 24 07:59:56 firewall pppd[9920]: local IP address 216.75.173.181
Jan 24 07:59:56 firewall pppd[9920]: remote IP address 64.39.160.16
Jan 24 08:00:01 firewall /USR/SBIN/CRON[12130]: (root) CMD
(/etc/keepalive_script)
Jan 24 08:00:01 firewall /USR/SBIN/CRON[26229]: (root) CMD
(`/etc/init.d/ntpdate start`)
Jan 24 08:00:01 firewall /USR/SBIN/CRON[25950]: (root) CMD (/etc/multicron-p)
Jan 24 03:00:07 firewall ntpdate[24300]: step time server 132.246.168.148
offset 0.357184 sec
Jan 24 08:00:07 firewall /USR/SBIN/CRON[20675]: (root) MAIL (mailed 19 bytes
of output but got status 0x0001 )
Jan 24 08:10:01 firewall /USR/SBIN/CRON[30855]: (root) CMD
(/etc/keepalive_script)
Jan 24 08:15:01 firewall /USR/SBIN/CRON[14604]: (root) CMD (/etc/multicron-p)
Jan 24 08:20:01 firewall /USR/SBIN/CRON[2738]: (root) CMD
(/etc/keepalive_script)
Jan 24 08:30:01 firewall /USR/SBIN/CRON[1570]: (root) CMD
(/etc/keepalive_script)
Jan 24 08:30:01 firewall /USR/SBIN/CRON[21929]: (root) CMD (/etc/multicron-p)
Jan 24 08:40:01 firewall /USR/SBIN/CRON[22420]: (root) CMD
(/etc/keepalive_script)
Jan 24 08:45:01 firewall /USR/SBIN/CRON[15576]: (root) CMD (/etc/multicron-p)
Jan 24 08:50:02 firewall /USR/SBIN/CRON[17687]: (root) CMD
(/etc/keepalive_script)
Jan 24 08:58:38 firewall pppd[9920]: No response to 10 echo-requests
Jan 24 08:58:38 firewall pppd[9920]: Serial link appears to be disconnected.
Jan 24 08:58:38 firewall pppd[9920]: Couldn't increase MTU to 1500.
Jan 24 08:58:38 firewall pppd[9920]: Couldn't increase MRU to 1500
Jan 24 08:58:48 firewall pppd[9920]: Connection terminated.
Jan 24 08:58:48 firewall pppd[9920]: Connect time 58.8 minutes.
Jan 24 08:58:48 firewall pppd[9920]: Sent 4236 bytes, received 8209 bytes.
Jan 24 08:58:48 firewall pppd[9920]: Doing disconnect
Jan 24 08:59:18 firewall pppd[9920]: Sending PADI
Jan 24 09:00:01 firewall /USR/SBIN/CRON[23934]: (root) CMD
(/etc/keepalive_script)
Jan 24 09:00:01 firewall /USR/SBIN/CRON[6709]: (root) CMD
(`/etc/init.d/ntpdate start`)
Jan 24 09:00:01 firewall /USR/SBIN/CRON[27318]: (root) CMD (/etc/multicron-p)
Jan 24 04:00:06 firewall ntpdate[18431]: sendto(132.246.168.148): Network is
unreachable
Jan 24 04:00:09 firewall last message repeated 3 times
Jan 24 04:00:10 firewall ntpdate[18431]: no server suitable for
synchronization found
Jan 24 09:00:10 firewall /USR/SBIN/CRON[14147]: (root) MAIL (mailed 19 bytes
of output but got status 0x0001 )
Jan 24 09:10:01 firewall /USR/SBIN/CRON[19360]: (root) CMD
(/etc/keepalive_script)
Jan 24 09:15:01 firewall /USR/SBIN/CRON[2952]: (root) CMD (/etc/multicron-p)
Jan 24 09:20:01 firewall /USR/SBIN/CRON[14049]: (root) CMD
(/etc/keepalive_script)
Jan 24 09:30:01 firewall /USR/SBIN/CRON[18832]: (root) CMD
(/etc/keepalive_script)
Jan 24 09:30:01 firewall /USR/SBIN/CRON[16748]: (root) CMD (/etc/multicron-p)
Jan 24 09:33:24 firewall pppd[9920]: Connecting PPPoE socket:
00:90:1a:40:44:2c 0000 eth0 0x807c2c8
Jan 24 09:33:24 firewall pppd[9920]: Couldn't get channel number: Transport
endpoint is not connected
Jan 24 09:33:24 firewall pppd[9920]: Doing disconnect
Jan 24 09:33:54 firewall pppd[9920]: Sending PADI
...
This fills my logs for a while until eventually the system stops trying.
The script "keepalive" is listed below:
if [ -z "`ip link | grep ppp0`" ]; then
/etc/init.d/networking restart
sleep 1
fi;
# If the DSL connection is up, keep it alive with ping.
if [ ! -z "`ip link | grep ppp0`" ]; then
ping -c 10 www.google.com 2>&1 > /dev/null &
fi
-- -------------------------------------------"Never tell me the odds!"-------- Ray Olszewski -- Han Solo Palo Alto, California, USA [EMAIL PROTECTED] -------------------------------------------------------------------------------
------------------------------------------------------- This SF.NET email is sponsored by: SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! http://www.vasoftware.com ------------------------------------------------------------------------ leaf-user mailing list: [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user SR FAQ: http://leaf-project.org/pub/doc/docmanager/docid_1891.html
