RV042 Connectivity Loss after upgrade to 1.3.13.02

[james.hicks]

After upgrading an RV042 from 1.3.12.19 to 1.3.13.02, it loses connectivity after several hours.

RV042 WAN1 is connected to cable modem (WAN2/DMZ is unused), configured to obtain IP Address.

Cold booting all devices, RV042, AP & Cable modem, re-established connectivity. Only to lose connectivity again several hours later.

Reset RV042 to factory default via reset switch and cleared log. Minimal configuration applied (domain name, hostname, ntp, WAN1 bw=8000).

Checked early this morning and no connectivity. Summary screen showed the same IP on WAN1 port. All appeared normal.

Clicked renew, and connectivity was re-established. Same IP as before.

The RV042, when running 1.3.12.19, would run for many months without a hiccup.

This is the log entry when I clicked 'renew':

Aug 18 04:12:40 2011

System   Log

DHCP   lease info: ip[68.202.x.x] mask[255.255.240.0] gw[68.202.x.x] host_name[starwars1rv042rtr]   domain_name[cfl.rr.com] server_name[NOT GIVEN] dns_server1[65.32.5.111]   dns_server2[65.32.5.112] ttl[64] mtu[576] renew_time[2011/8/18 05:05:10]   expire_time[2011/8/18 05:12:40]

Any ideas what might be the issue?

Clicking 'renew' did re-establish connectivity, but I'm not sure if it's a dhcp issue or not.

If I read the log entry correctly, the router has the IP lease for an hour, which means it renews several times successfully. Connectivity is down in the morning after several hours go by without use. I'm not sure if that has anything to do with it. Again, nothing changed but the code and the RV042 ran without problems for months with the 1.3.12.19 code.

jay

[jasbryan]

Jay,

I would downgrade back to earlier firmware version and test. Just to verify that it's code and not configuration. We have limited configuration we can make on the wan connection. Before you downgrade try this; plug your PC directly into your cable modem and get an public ip address, now plug PC directly into RV0xx router without plugging router into cable modem. Now clone your PC Mac address, now plug router back into modem and reboot router. Make sure MTU is set to AUTO after completing these steps, and problem still persist please downgrade back to earlier firmware for re-testing.

While on latest firmware, try and capture some logging information. If you downgrade to 1.3.12.19 code and it works perfectly we would need to have it documented and have you call in for support so can create case and escalate to our product engineers for possible bug in 1.3.13.02

Thanks

Jasbryan

Cisco Support Engineer

.:|:.:|:.

[james.hicks]

Thank you Jasbryan for the quick reply. I'll perform testing per your suggestion and provide appropriate feedback.

Regards,

Jay

[james.hicks]

Jasbryan,

Please review attached log file (pdf). I checked @ 10 hours after I manually 'renewed' the IP address from the summary page early this morning. This afternoon, the RV042 was on-line and connectivity was normal, not off-line as I thought it might be.

Instead of disturbing the router, I reviewed the log file a little closer and it appears the DHCP process was confused by erroneous time. I don't have the logs immediately following the upgrade a few days ago, but I suspect time was incorrect then as well.

The router restart Aug 17 09:52:36 2011, yesterday, was the manual reset (rear panel button). Note system/log time when the system comes back up after the reset: "Dec 31 19:00:01 2002".

Then, note the calculated expire_time: "2003/1/1 06:34:47", clearly in the past, and probably the cause of the problem. The system clock is corrected after it gets it's lease, when it can get to a NTP server (but too late for the current lease).

Early this morning, after the manual 'renew' from the summary page, the expire_time is correctly calculated in the future: 2011/8/18 05:12:40 (calculated off the present/correct time).

The actual expire_time/lease_renew occur on time: Aug 18 05:12:51 2011

I expect the dhcp process to work properly again later today and should see a log entry @ 2011/8/18 17:12:42.

2nd automatic renew occured at: Aug 18 17:12:51 2011

Conclusion; the dhcp process was confused by incorrect system clock which it set the lease expire time in the past.

Regards,

Jay

[jasbryan]

James,

Is this logs from the existing firmware 1.3.13.02 or did we downgrade back to 1.3.12.19 and capture the logs from it.

Also what is your Cisco User ID?

Jasbryan

Cisco Support Engineer

.:|:.:|:.

[james.hicks]

The log is from v1.3.13.02. The router is passing traffic with no problems since ntp updated the internal clock. Looking at the log, the dhcp process renew and expire times were in the past probably confusing the process. As soon as ntp updated the clock AND I manually renewed the IP all has been well. Do you think there's an issue with the code? jay (james.hicks)