AC68U - dual WAN stability: ISP's DHCP did not function properly

[texashoser]

Got an email back from Asus technical support. They said the latest release should fix the problem. I told them I was already running the latest firmware. Now, crickets...

 

[texashoser]

Asus has closed all my open tickets on the issue. Don't know if engineering ever investigated the issue. Anyway, my opinion of Asus is that they suck.

 

[texashoser]

Watchdog still broken in the 4376 release.

I did get a response from Gary, though. They are looking into it...

 

[texashoser]

To Asus' credit, I'm finally working with an engineer to hopefully solve the bug in watchdog feature. Fingers crossed.

 

[texashoser]

In the last week I've finally been able to make some progress with their engineering team. They've stated that the PPPoE error on my ATT static IP connection is a bug. They tried to claim that the "DHCP did not function correctly error" was either the watchdog host or my ISP dropping the connection due to excessive ICMP packets. I told them if the connection were being dropped due to that, I shouldn't be seeing DHCP errors, just a plain disconnect. I also pointed out that if their theory was correct, if I turned off the watchdog feature and pinged the same watchdog host at the same interval from a computer behind the router, then I should see the router behave exactly the same. It doesn't, of course. The connection is never dropped and the logs indicate this.

So... They are trying to recreate the problem. Have no idea how long it will take.

 

[texashoser]

FYI, Asus engineering now treating both errors as bugs, are working to duplicate the problem, and will send me a BETA release when they've made some progress.

 

[Charlie Morrison]

I was glad to find this thread. We're having trouble with the Dual Wan with watchdog setting on our RT-AC68R, and we're seeing the same thing in our logs. We have TWC Business Class on the Primary WAN, and Verizon Wireless 4G modem on the Secondary WAN (on a LAN port, not USB). We are currently at firmware 3.0.0.4.378_4585. The primary WAN is a dedicated IP, but the Verizon 4G modem is dynamic IP.

Will the router fail over and fail back without the watchdog? It has a failback check that is still enabled, and failover and failback detection times still enabled.

 

[texashoser]

Yes, the router should fail over to the secondary connection if watchdog is disabled, but only if the ethernet link goes down - which is the most unlikely scenario, unfortunately. The Time Warner connection may go down, have problems, etc, but the modem itself is not likely to fail, etc, in order to cause a link down situation.

And even in this rare case, it takes a long time for the fail over and fall back.

I'm running a beta load that supposedly fixes the PPPoE errors when using a WAN connection with static IP addresses when watchdog is enabled. In my case, though, that static IP connection is the secondary one, not primary, so this beta load might actually work for you. They are still looking into what is causing the "DHCP did not function properly" errors when using a DHCP WAN connection with watchdog enabled.

 

[Puppa]

Just saw this thread...I had this exact scenario happen with my Time Warner cable connection (but non-PPoE!) on an RT-AC3200 last week. Twice, both with .9135 firmware. This was not apparently due to the known memory leak problem with that FW, as RAM was under 40% used. "DHCP did not function correctly error"

Dang...considering whether I should abandon ASUS...I have had serious issues with every consumer wireless router, so I may just be jumping from one frying pan to another frying pan...

 

[texashoser]

Just saw this thread...I had this exact scenario happen with my Time Warner cable connection (but non-PPoE!) on an RT-AC3200 last week. Twice, both with .9135 firmware. This was not apparently due to the known memory leak problem with that FW, as RAM was under 40% used. "DHCP did not function correctly error"

Dang...considering whether I should abandon ASUS...I have had serious issues with every consumer wireless router, so I may just be jumping from one frying pan to another frying pan...

You typically should only being seeing the DHCP error when you have dual WAN turned on with the watchdog feature enabled.

 

[Puppa]

Dual WAN is not setup, but it still happens. Happened just past midnight last night again. Syslog also showed evidence of RAM corruption (garbage characters in middle of string)

Sent from my SM-N910V using Tapatalk

 

[texashoser]

Dual WAN is not setup, but it still happens. Happened just past midnight last night again. Syslog also showed evidence of RAM corruption (garbage characters in middle of string)

Interesting. I never see the DHCP error on my TW connection with watchdog enabled. In any case, seems to be a bug which they haven't figured out yet.

They claimed it was Time Warner resetting my connection or filtering ICMP packets that was causing the DHCP problem. Except that neither of those conditions would cause the router to make a DHCP request. Only a physical link up/down situation or the lease time expiring (or some percentage of the lease expiration time) would do that. Not only that, but I pinged the watchdog host (in my case, www.google.com) every second from not only a box behind the router but from the router itself and a) never dropped a packet and b) with the watchdog feature turned off, never got a DHCP error.

Again, they claim to have the PPPoE error (which in my case I'm not even using PPPoE with ATT) fixed but are still looking into the DHCP error/watchdog problem.

 

[Puppa]

If this RAM corruption is a thing, it may be quite difficult to track down the root cause. The overwrite mid-printk was definitely binary data, not ascii.

Sent from my SM-N910V using Tapatalk

 

[Year10K]

Having the same issue here with Dual WAN. Primary is Comcast DHCP, secondary is U-Verse DHCP. Keep getting the DHCP error. I recently got U-verse as backup, and was about to order TP-LINK TL-R470T+ for failover. Luckily (or so I thought), discovered that my Asus RT-AC68U has this feature. I am using latest firmware, 3.0.0.4.378_9313. Bummer that this great feature is broken and Asus doesn't seem interested in fixing it. Paid a lot of money for this router, and this would have made it worth the expense. Hoping to avoid having to add the TP-LINK. More expense, clutter, and another item to maintain (firmware, bugs, etc.).
Texashoser, impressed that you've pursued ASUS for almost 1.5 years on this issue...Unbelievable that they haven't fixed in 1.5 years!

 

[Year10K]

I tried to use the failover without the watch dog, because usually when I lose Comcast the actual link goes down, so I thought that should work. Unfortunately, that does not appear to work either. I lost my Comcast link, such that only one modem light was on and blinking, and the ASUS web console clearly showed the primary WAN as "Disconnected" and Secondary as "Standby". It never switched over. I had to physically pull the ethernet cable from the Comcast modem, and then it switched. I tried plugging the ethernet cable back into the Primary WAN while it was still down, and the ASUS put the secondary WAN to Standby, and the Primary WAN showed as Disconnected, so lost the internet connection.

So looks like the ASUS is only good for manual failover (disconnect cable) and manual failback (re-connect cable). I guess I'll be getting the TP-LINK after all, as my Comcast goes down frequently so I really need something to do switching automatically and reliably.

 

[texashoser]

Year10K, it did take a while to get them working on it, but they have been of late. They've given me two beta loads, the first supposedly getting rid of the PPPoE error for WAN connections using static IP's.

They thought they had fixed the DHCP error by changing up the part of the watchdog code that employs ICMP packets (they were using a different ping than what's used at the shell level), but that didn't work. So, they are still working on it.

 

[texashoser]

Update: The latest version of beta FW provided by Asus does indeed seem to have fixed the issue (for the most part) of getting WAN disconnects via the DHCP did function properly bug. After upgrading, I saw the error about five hours later but since then (five days ago) haven't seen any disconnects. Crossing my fingers that this is finally fixed (beta load is also supposed to have fixed the PPPoE errors for static IP WAN connections.

Remember, these errors only happen when the watchdog function is enabled when dual WAN is turned on in fail over mode.

 

[texashoser]

Nine days and counting with no errors!

 

[texashoser]

Just wanted to close this thread up. While dual WAN seems still to have some issues, most of the problems have been addressed by engineering and overall the stability is now vastly improved. Watchdog issue was, for the most part, fixed in 3.0.0.4.380.1031 released back in December.