Hello, I am experiencing a strange, intermittent problem with pppd. I am running pppd version 2.4.4 under Debian "Lenny" on an old Dell Optiplex GX260 using kernel 2.6.26-2-686. I have a Calcomp A600 3G Wireless modem dialing a broadband wireless carrier. The system is configured as a linux router running an openvpn client over the ppp0 connection, with dnsmasq providing DHCP clients on eth0 with IP configuration and ntp providing a clock source. If I leave the router up and running, after 12 hours the ISP disconnects the line. I can avoid this by sending a SIGHUP every few hours. I of course have pppd running with the persist option, and usually pppd will dutifully hang up the modem and start the chat session over again to obtain connectivity. Frequently, however, pppd will simply hang rather than restarting the connection. Subsequent SIGTERM or SIGHUP signals will do nothing whatsoever. The only way (of which I know) to release the terminal (/cev/ttyACM0) is to send a SIGKILL. After doing so, /dev/ttyACM0 is released, and I can use a utility such as minicom to access the modem. I can dial out and get a carrier. I can see the ppp authentication request from the ISP. What I cannot do, however, is get pppd to come back up. If I start pppd again, it opens the /dev/ttyACM0 device and creates the lockfile, but it does not attempt to run chat or evidently do anything else other than refresh the lockfile if it is stale. My only option at that point is to reboot the host. A warm reboot allows the router to call out again. Rebooting the router is a rather Draconian approach, not to mention the fact it takes a little bit of time. The problem by no means happens every time, but it does happen very frequently, perhaps one restart out of three. Is there some way other than sending a SIGHUP to get pppd to hang up gracefully and redial the line? Is there something I can do to get pppd to start working properly when it hangs, short of rebooting the router? Here is a snapshot from /var/log/syslog during one of the events: Jan 19 19:35:29 Cricket pppd[2783]: Hangup (SIGHUP) Jan 19 19:35:29 Cricket pppd[2783]: Connect time 152.1 minutes. Jan 19 19:35:29 Cricket pppd[2783]: Sent 1551485 bytes, received 1648658 bytes. Jan 19 19:35:29 Cricket pppd[2783]: Script /etc/ppp/ip-down started (pid 10026) Jan 19 19:35:29 Cricket pppd[2783]: sent [LCP TermReq id=0xe "User request"] Jan 19 19:35:29 Cricket pppd[2783]: Script /etc/ppp/ip-down finished (pid 10026), status = 0x0 Jan 19 19:35:32 Cricket pppd[2783]: sent [LCP TermReq id=0xf "User request"] Jan 19 19:35:35 Cricket pppd[2783]: Connection terminated. Jan 19 19:39:00 Cricket pppd[2783]: ioctl(TIOCSETD, N_TTY): Interrupted system call (line 571) Jan 19 19:39:01 Cricket pppd[2783]: tcsetattr: Interrupted system call (line 1010) Jan 19 19:39:01 Cricket pppd[2783]: Hangup (SIGHUP) Jan 19 19:39:01 Cricket pppd[2783]: Modem hangup Jan 19 19:39:02 Cricket pppd[2783]: Hangup (SIGHUP) Then nothing else. After I kill the process and restart pppd, I get only a single line: Jan 19 20:11:45 Cricket pppd[18566]: pppd 2.4.4 started by root, uid 0 Another time the log had the same results except it quit logging after the ioctl(TIOCSETD, N_TTY) line. I call pppd with `pppd call cricket` Active options in the /etc/ppp/options file: asyncmap 0 usehostname noipdefault usepeerdns defaultroute auth local lock hide-password debug lcp-echo-interval 30 lcp-echo-failure 4 noipx Contents of /etc/ppp/peers/cricket: user "(830)388-4911@xxxxxxxxxxxxx" connect "/usr/sbin/chat -v -f /etc/ppp/chatfile" /dev/ttyACM0 noipdefault usepeerdns defaultroute persist noauth Contents of /etc/ppp/chatfile ABORT "NO CARRIER" ABORT "NO DIALTONE" ABORT "ERROR" ABORT "NO ANSWER" ABORT "BUSY" "" "AT" OK "ATDT#777" CONNECT -- To unsubscribe from this list: send the line "unsubscribe linux-ppp" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
