Posted: Sun Jan 01, 2012 7:52 Post subject: WDS WAtchdog
Alright so i have Linksys E3000 which is my main router and 2 Netgear WNDR3300in WDS which are working perfectly now. I enabled WDS watchdog on both netgear and set it to ping "192.168.1.1" which is the E3000's IP address, so it should reboot if it looses connection.
I have questions which i'm not to sure about.
1.) Should i enable WDS watchdog on the Main Router as well and have it ping one or both netgear?
A few times i was forced to reboot the main router just so the other one would connect which is why i'm thinking its a good idea to ping the netgear's from the main.
2.) The interval is set to 1000 seconds, does that mean the router will reboot every 1000 seconds if it can't ping the main router?
I was able to figure out the answer to number 2 which is, it pings every 1000 seconds and if no connection it reboots itself.
Question 1 still applies, and i'm not to sure if the Main need watchdog enabled or just scheduled to reboot every x day.
Did a few more test and if one of the netgear are not not connected via WDS the main router still shows a percentage in signal strength and i think this is why the connection is not established until the main router is rebooted.
Like you, I too have several WNDR3300 routers; three of them are connected in repeater bridge mode to the main router using the wl0 radio (5 GHz - N only). I prefer to not reboot using the WDS watchdog. Instead, I have a startup script like the one below on the three repeater bridges to take down wl0 and bring it up again if they are unable to ping the main router. I find this is sufficient to keep the routers "talking" to each other.
Code:
#!/bin/sh
while : ; do
count=`ping -c3 192.168.1.1 2>/dev/null | awk '/packets received/ {print $4}'`
if [ $count -eq 0 ]; then
wl -i eth2 down ; wl -i eth2 up
else
echo "success" >/dev/null
fi
sleep 600
done
Regarding your first question, I do not reboot the routers periodically. Their uptimes range from 31-48 days. There's no noticeable decrease in performance, as measured by occasionally using speedtest, and streaming HD video daily.
Truth be told, I don't know if the routers have ever lost connectivity "spontaneously". But I know the script works because I have intentionally unplugged the main router to test the script.
jebise101 wrote:
I was able to figure out the answer to number 2 which is, it pings every 1000 seconds and if no connection it reboots itself.
Question 1 still applies, and i'm not to sure if the Main need watchdog enabled or just scheduled to reboot every x day.
Did a few more test and if one of the netgear are not not connected via WDS the main router still shows a percentage in signal strength and i think this is why the connection is not established until the main router is rebooted.
Thanks this will actually help. It seems i will be returning the E3000 so will having to use Client Bridge rather then WDS on the 3300's because i will keep the WNDR3700. I don't really care if 3300's reboot or not as long as they keep the connection alive.
Your script will help if watchdog does not work in Client Bridge mode. This will be of great use on the main router as well, as i would hate to bring down the entire network because one of the 3300's stopped taking. The 5GHz band only has other routers connected to it so just restart wl0 mean no one else gets disconnected.
Yes i agree the 3300's are very solid, mine where setup in WDS until i decided to upgrade and went down maybe twice in the entire year.
I was initially using WDS Watchdog but loosing connection shouldn't reboot the router for a longer startup. So I was searching for a solution to restart the WDS Radio and came to this post and help me come up with my solution to share. I tested this code on a WDS Network with three E3000s using K3.10 Build 35531.
<KONG> posted these commands...
eth1 radio off command "startservice radio_off_0 -f"
eth1 radio on command "startservice radio_on_0 -f"
eth2 radio off command "startservice radio_off_1 -f"
eth2 radio on command "startservice radio_on_1 -f"