Re: Interesting WAN failure problem

Brian Bergin · ‎12-17-2010

So we have hundreds of locations where we've recommended RV082's since they were released and have had the occasional issue with WAN failover or removal not working when a WAN connection was “down” but last week we think we discovered why RV082’s don’t always drop “dead” connections when they’re down. We had a customer with a both a cable modem and DSL modem in load balance mode and while the cable modem could ping and could get enough traffic to do a DNS lookup it could never load a HTML page. Even connected directly to a laptop I could ping with great times to anyplace I tried but loading a web site like Yahoo or Google failed 100% of the time. I took our v3 RV082 to the customer and found the same problems, the RV082 simply failed to realize that while ICMP worked, HTML didn’t.

I submit we need a way to have an RV082 pull HTML code from some page somewhere perhaps on a cisco.com site, simple HTML will do, not even <HTML>Hello does this work?</HTML> wouldn’t load when pings worked, and have the RV0xx’s check HTML as well as ICMP pings to hosts to determine if a WAN port is bad. I’m betting this is the reason other customers have had to actually unplug a WAN Port to get failover to work right on RV082’s in the past.

Kurt Schumacher · ‎12-19-2010

Well, does remind me to the WAN fail-over issues I tried to report before - with no result until now.

Had no experience with the RV0xx-V1/-V2 before, but I'm very disappointed about the RV082-V3. As-is, certianly not a router I wuold promote to friends, customers, or the worst enemy -. and this is not only because of the reported and non-addressed issues after a very long time now. Might be nice as a no-name low cost device, but certainly now with a Cisco brand.

Either the box get a massuve bottom-up firme upgrade (including fixing specs...), or it's a no-go in my opinion.

-Kurt.

Te-Kai Liu · ‎12-19-2010

The WAN failover issue has been fixed in a new firmware that has not gone through the testing. Let us know if you want to test drive it.

Brian Bergin · ‎12-19-2010

Absolutely I want to test drive it on our v3 ET units, but the question is will Cisco fix the fail over issue in the hundreds of v2 appliances we have out there? When RV082's work well they're wonderful. I don't share Kurt's negativity about the units, but for an ~8 year old product there are enough issues that haven't been worked out to question our continued loyalty to the model.

Honestly, the best shot at replacing this was the SA500 series but that's fallen short of expectations too. Cisco will contiue to fail in the SB marketplace until it fixes its internal proceedures for product design AND implementation. It's literally that simple if Cisco wants to become a real player in the SB marketplace.

Te-Kai Liu · ‎12-19-2010

will Cisco fix the fail over issue in the hundreds of v2 appliances we have out there?

If we can duplicate and root cause the "fail over" issue you mentioned, the issue will likely be fixed in the next maintenance release.

Kurt Schumacher · ‎12-20-2010

tekliu wrote:

The WAN failover issue has been fixed in a new firmware that has not gone through the testing. Let us know if you want to test drive it.

Sure, we're here to look into the potentially fixed - just make it available to us please.

Brian,

None of the mainstream broadband providers here in Switzerland is offering IPv6, too. A few very small engineering-driven ones offer IPv6 address ranges however.

We have just been invited to a Cisco SB IPv6 training on January 19 - I guess the teacher will have a hard time...

-Kurt.

Brian Bergin · ‎12-20-2010

None of the mainstream broadband providers here in Switzerland is offering IPv6, too. A few very small engineering-driven ones offer IPv6 address ranges however.

We have just been invited to a Cisco SB IPv6 training on January 19 - I guess the teacher will have a hard time...

The entire problem, IMHO, with both of Cisco's SB divisions is a lack of understanding of what the real smal business world needs and this lack of understanding doesn’t start, I'm betting, in the SB division, it starts several pay grades above them. 99.99% of the real world SB users have ZERO need for IPv6 on the public side in 2010. 95% won't by 2015 and the 5% that do then will not be the ones using RV082’s today.

My guess is the SB divisions are getting pressure from very high up in the chain of command to implement IPv6 so Cisco can claim full IPv6 support across their entire product line, but this is at the expense of fixing problems that actually affect us all today. That’s a shame and it will continue to leave Cisco on the Small Business sideline as a minor player with little to no influence on the market segment. Eventually Cisco’s CEO/President/etc… will either figure this out and will blame it on the SB divisions or, as suspect will be the case, Cisco’s investors will figure it out and find a new CEO/President/etc… who will let the folks who run the SB divisions create products we need NOT products that look good on paper.

How many bridges did I just burn? I hope not too many, but I tell it like I see it. The PMs in the SB groups needs to stand up for their customer base. Give me 1 good reason why I should be excited about IPv6 in an RV082 and I’ll shut up, but for now I couldn’t get an IPv6 address from any of our broadband providers if I begged. The ONLY thing IPv6 works good for is on Win7 home networks for creating sharing groups, but that’s it and that doesn’t require a router to support it as Win7 does its own config of IPv6 in those cases. Come on Cisco, let’s work on the real problems, forget IPv6, it’s useless to anyone who NEEDS a Cisco SB router today.

Te-Kai Liu · ‎12-19-2010

I would clarify that the V3 failover issue that was fixed is that the failover time is longer than that of V2 by about 20~30 sec.

Brian Bergin · ‎12-19-2010

So the issue here that the RV0xx actually being able to detect if HTML and other real IP traffic, not ICMP, hasn't been fixed? If that's the case the question becomes quickly, will it get fixed? If not, then the RV0xx should be retired and Cisco's entire focus should shift to fixing the SA500's load balancing issues (it doesn't actually load balance, it round robin's the WAN connections) as well as it's ~2 minute failover time if you're not in load balance mode (among other issues).

Honestly, I had high hopes for the RV's as they've been our staple for many years, but they've fallen well behind the times. I know the v3 units are a "simple" rev of the v2 units for new CPUs, but this was the time to go back and fix what has been missing, but I guess that's just not happening. That's a shame, they have such potential for a sub $300 unit. I again stress to Cisco that feature design and implementation and not product physical quality is why Cisco has and will continue to fail in the small biz world.

Te-Kai Liu · ‎12-19-2010

Adding html to the network service detection module is something that can be considered. But a fundamental question is under what circumstance will ping going through but html failing to pass?

On the V3 side, we are working on adding IPv6 support to V3 in the next Maintenance Release.

Brian Bergin · ‎12-19-2010

That's a question for someone with more knowledge of TCP/IP and ICMP than me, but I can tell you it happens all the time. Pings will pass and for that matter DNS will resolve but larger amounts of traffic, say HTML or IPSec fail.

As for IPv6, let's get the rest of the design flaws worked out before you introduce IPv6, something unused by the vast majority of your target audience. I'd love to see a survey of Cisco SBs to see how many even know what IPv6 is let alone how and when to implement it.

To date, I've yet to have a broadband provider offer an IPv6 addy to a customer so while you might be planning for the future, by the time it trickles down to the SB broadband connection the v3 units will have long since been retired.

Anyone else out there have a broadband provider offering IPv6 addy's to customers?

ddiep · ‎12-22-2010

Brian at et,

I 'd just posted latest firmware version (4.0.1.2) here (Latest firmware (4.0.0.7 + 4.0.1.2) as of 12/22/2010 ). This fw had *not* gone thru verification. It addressed and fixed a few issues that raised from this forum. Use it however you seemed fit. Let me know if any other issue found.

Best regards and happy holiday.