Hello, there's something else in the logs I forgot to mention. This happends just before the connection is drop: Apr 21 22:36:05 gregg pppd[1099]: sent [LCP EchoRep id=0xc5 magic=0x58896215 1c 06 02 2d] Apr 21 22:36:06 gregg pppd[1099]: sent [LCP EchoReq id=0xd magic=0x58896215] Apr 21 22:36:06 gregg pppd[1099]: rcvd [LCP EchoRep id=0xd magic=0xaefddb9f] 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 Apr 21 22:36:19 gregg kernel: NETDEV WATCHDOG: eth0: transmit timed out Apr 21 22:36:19 gregg kernel: eth0: Tx timed out, cable problem? TSR=0x16, ISR=0x0, t=22. Apr 21 22:36:20 gregg kernel: NETDEV WATCHDOG: eth0: transmit timed out Apr 21 22:36:20 gregg kernel: eth0: Tx timed out, cable problem? TSR=0x16, ISR=0x0, t=25. I guess it is because the ISP peer has drop and the LCP echos are not sent anymore? But why is there a timeout in eth0? Shall the timeout be in ppp0? Or is there anything wrong with eth0? Thanks Pep Serrano.