PPPoE connection dropping intermittently / WAN interface reset
-
Hi all,
I've recently purchased an SG-3100 which is terminating my VDSL connection via a Draytek Vigor 130 modem over PPPoE.
This setup has been rock solid since December 1st until today. Four times today my PPPoE connection has dropped, however my VDSL modem has maintained sync.
Normally, when the VDSL modem loses sync and the PPP connection resets, you see these messages in ppp.log:
Nov 30 19:40:55 pfsense ppp: [wan] REDACTED -> REDACTED Dec 1 14:18:58 pfsense ppp: [wan_link0] LCP: no reply to 1 echo request(s) Dec 1 14:19:08 pfsense ppp: [wan_link0] LCP: no reply to 2 echo request(s) Dec 1 14:19:18 pfsense ppp: [wan_link0] LCP: no reply to 3 echo request(s) Dec 1 14:19:28 pfsense ppp: [wan_link0] LCP: no reply to 4 echo request(s) Dec 1 14:19:38 pfsense ppp: [wan_link0] LCP: no reply to 5 echo request(s) Dec 1 14:19:38 pfsense ppp: [wan_link0] LCP: peer not responding to echo requests Dec 1 14:19:38 pfsense ppp: [wan_link0] LCP: state change Opened --> Stopping Dec 1 14:19:38 pfsense ppp: [wan_link0] Link: Leave bundle "wan" Dec 1 14:19:38 pfsense ppp: [wan] Bundle: Status update: up 0 links, total bandwidth 9600 bps Dec 1 14:19:38 pfsense ppp: [wan] IPCP: Close event
The logs for the disconnects today give no indication as to why the link dropped, all I get is the following:
Dec 1 14:30:07 pfsense ppp: [wan] REDACTED -> REDACTED Dec 1 14:30:07 pfsense ppp: [wan_link0] rec'd unexpected protocol IP Dec 9 11:27:55 pfsense ppp: Multi-link PPP daemon for FreeBSD Dec 9 11:27:55 pfsense ppp: Dec 9 11:27:55 pfsense ppp: process 67445 started, version 5.8 (nobody@pfSense_factory-v2_4_1_armv6-pfSense_factory-v2_4_1-job-01 16:39 21-Oct-2017) Dec 9 11:27:55 pfsense ppp: waiting for process 98028 to die... Dec 9 11:27:55 pfsense ppp: caught fatal signal TERM Dec 9 11:27:55 pfsense ppp: [wan] IFACE: Close event Dec 9 11:27:55 pfsense ppp: [wan] IPCP: Close event Dec 9 11:27:55 pfsense ppp: [wan] IPCP: state change Opened --> Closing
When I look into system.log, it says the interface goes down & up:
Dec 9 11:27:51 pfsense check_reload_status: Linkup starting mvneta2 Dec 9 11:27:51 pfsense kernel: mvneta2: link state changed to DOWN Dec 9 11:27:52 pfsense check_reload_status: Reloading filter Dec 9 11:27:54 pfsense check_reload_status: Linkup starting mvneta2 Dec 9 11:27:54 pfsense kernel: mvneta2: link state changed to UP Dec 9 11:27:55 pfsense ppp: Multi-link PPP daemon for FreeBSD
The SG-3100 is connected to the modem by a brand new 0.5m Cat6a cable, in a locked closet - so I don't think it's the cable, but I will get one out to test.
Has anyone else experience similar, or could point to why this might happen, other than a faulty cable?
Thanks,
Verta. -
Has anyone else experience similar, or could point to why this might happen, other than a faulty cable?
If it worked fine for a few days I would guess your provider has some issues at the moment.
-
Has anyone else experience similar, or could point to why this might happen, other than a faulty cable?
If it worked fine for a few days I would guess your provider has some issues at the moment.
Thanks for the response. I've dug out a new cable and noticed that whenever I move the cable, the port on the SG-3100 goes down, whetever it's connected to the modem or a switch - maybe a faulty port? I'll drop Netgate a line.
-
You might try the cable first.
-
I'm four cables in and getting some interesting results. Any cable that has a shielded RJ45 connector, the port will drop when the cable is moved. Any unshielded RJ45 connector and the port doesn't drop. I'm reporting back to the reseller I bought the unit from to see if they think it's faulty or not.
Thanks for the assistance.
-
One thing about shielded cables. They're supposed to be grounded at one and only one point. If they're not grounded, the shield is ineffective. If grounded at more than one point, ground loops may occur.
However, given that just moving the cable causes failure, it's likely a poor connection somewhere.Rule of thumb, when something fails, cables and connectors are the likely suspects.