Internet slowing down terribly w/ 378.55

[kobari]

Hello All,

I have Asus AC-68u w/ 378.55. And, I restart my router every night with a cronjob. But, during the day, what I am experiencing is that after some certain amount of time (couple of hours), Internet is slowing down to a grinding halt. Google ping times are slowing down x3 times or so, bandwidth is dropping to between 1/5th and 1/6th of normal.

And, I have one at work (same router w/ 378.55), and that is even worse where I loose the Internet connection completely (cannot ping google from connected devices or even from the router via ssh).

BTW, internal network is completely fine, I can connect to the router or anything on LAN. I am just having problems w/ the internet connection.

And, weird thing is that, restarting the router is restoring the speeds / connectivity immediately, literally right after a restart, speed is resumed to normal values.

I tried looking at syslog, but not seeing anything valuable. And, I am quite desperate at this point, since I need to keep restarting the router every couple of hours...

As a side note, I am not running anything special on the router like IP traffic logging, or DLNA, or advanced QoS or torrent or anything that might be a resource hog.

Can you please recommend what I should be looking at and/or how can diagnose this issue?

Thanks a ton.

 

[Zirescu]

Does your connection speeds get restored if you just reboot the modem instead? Ran into an issue at a customer's site where the modem was going sideways, a reboot of the modem fixed things for a period of time. Eventually had the modem replaced and no more issues. Their ISP had done some firmware updates on the original modem, so suspect that it had caused part of the problem, might be worth checking.

 

[mstombs]

What is the wan type? If dhcp does the slowdown correspond to failure to renew the lease? If so you can use a ping script to renew the dhcp or restart the wan if needed. I have similar problem with dhcp wan but wan lease is 7 days so nothing like the same issue!

 

[kobari]

Hi,

Wan Connection Type is Static IP and Wan Type is set to WAN.

For LAN DHCP, TTL is 1 day.

Basically, after a while we see: "Primary WAN: Disconnected"

Btw, how can we restart the WAN without rebooting the router?

What is the wan type? If dhcp does the slowdown correspond to failure to renew the lease? If so you can use a ping script to renew the dhcp or restart the wan if needed. I have similar problem with dhcp wan but wan lease is 7 days so nothing like the same issue!

 

[kobari]

Wrote more scripts to catch when exactly this issue starts, and saw this in syslog.log right at the point we lost the connection:

Aug 12 15:32:40 WAN Connection: Detected that the WAN Connection Type was PPPoE. But the PPPoE Setting was not complete.
Aug 12 15:32:40 stop_nat_rules: apply the redirect_rules!
Aug 12 15:33:29 dnsmasq-dhcp[507]: DHCPREQUEST(br0) 192.168.1.246 nn:nn:nn:nn:nn:nn <REPLACED_MAC>
Aug 12 15:33:29 dnsmasq-dhcp[507]: DHCPACK(br0) 192.168.1.246 nn:nn:nn:nn:nn:nn <REPLACED_MAC>
Aug 12 15:33:33 kernel: TCP: Possible SYN flooding on port 18017. Sending cookies.
Aug 12 15:33:33 kernel: TCP: Possible SYN flooding on port 18017. Sending cookies.
Aug 12 15:33:40 kernel: TCP: Possible SYN flooding on port 18017. Sending cookies.
Aug 12 15:33:40 kernel: TCP: Possible SYN flooding on port 18017. Sending cookies.
Aug 12 15:33:40 kernel: TCP: Possible SYN flooding on port 18017. Sending cookies.
Aug 12 15:33:40 kernel: TCP: Possible SYN flooding on port 18017. Sending cookies.
Aug 12 15:33:40 kernel: TCP: Possible SYN flooding on port 18017. Sending cookies.
Aug 12 15:33:40 kernel: TCP: Possible SYN flooding on port 18017. Sending cookies.
Aug 12 15:33:40 kernel: TCP: Possible SYN flooding on port 18017. Sending cookies.
Aug 12 15:33:40 kernel: TCP: Possible SYN flooding on port 18017. Sending cookies.
Aug 12 15:33:40 kernel: TCP: Possible SYN flooding on port 18017. Sending cookies.
Aug 12 15:33:40 kernel: TCP: Possible SYN flooding on port 18017. Sending cookies.

Also, have been monitoring temperatures, cpu load and memory and see no issue there...

 

[mstombs]

That reference to PPPoE shows that it is the router shutting down the WAN the redirect is for all http requests to be diverted to the router. With static wan IP that just looks wrong!

Script suggestions to restart WAN in this thread

WAN connect and disconnect using script

 

[kobari]

That is exactly what I was thinking. Any idea what is going on here and / or how I can figure out what the issue is?

Thanks.

 

[mstombs]

That error message comes from the asus program wanduck, it looks like the reference to pppoe is bogus - the program just assumes if not dhcp must be pppoe! You are not the only one with this problem, but seems more common with dualwan setups.

https://github.com/RMerl/asuswrt-me...f5b1d9a/release/src/router/rc/wanduck.c#L1555

But to get to line of code it has to believe that CASE_DHCPFAIL which still doesn't feel right with static wan connection, but probably comes from around here:-

https://github.com/RMerl/asuswrt-me...cf5b1d9a/release/src/router/rc/wanduck.c#L760

Note the error message from similar code location for dhcp connections is pretty common:-

Jul 27 22:44:51 WAN Connection: ISP's DHCP did not function properly.

which is a non-fatal error, occurs on every wan reconnect for me.

 

[kobari]

Thanks for the details.

One thing to note is that, this started happening after I upgraded to 378.55, so was wondering if this is related to that upgrade.

It got to a really bad state where I was getting disconnects every hour or so, then I switched to PPPoE and put the modem in bridge mode, but this time I started getting:

"WAN Connection: Fail to connect with some issues"

without any other explanation. And, this was even more frequent.

This got to such a bad state that, I had to remove my ASus and had to switch to Airport Extreme. And, after I switched to Airport Extreme, we had no issues whatsoever. So, I suspect, there is a problem with this firmware???

Once again, any help is appreciated.

Thanks.

 

[Traste]

+1
I have also noticed a significant slowdown on AC68u, and after a reboot RAM goes up immediately to 88% usage, and CPU temp is at 83C.

I am not sure about temp, if that is new or unusual, but i think i would have noticed. Prior to 378.55 i have not seen these behaviours.

BTW. the reason i noticed a real difference was that game servers started kicking me for high latency/high ping times, a problem i never had before.

 

[Traste]

+1
I have also noticed a significant slowdown on AC68u, and after a reboot RAM goes up immediately to 88% usage, and CPU temp is at 83C.

I am not sure about temp, if that is new or unusual, but i think i would have noticed. Prior to 378.55 i have not seen these behaviours.

I just downgraded to 378.54_2 and memory used went down to 33% immediately after reeboot, then slowly grew to around 46%.
Connections seems much perkier, and i will wait a while to see what the CPU temp says after an hour or so.

Granted, i did not follow any instructions when i did the last upgrade, just uploaded, so i probably missed something i should have done.