Dear Forum.
Occasionally we experience problems with our internet connection. We are tracing this problem for quite some time now.
Due to the fact that we are little far away from the DSLAM we do not have the fastest DSL. We are trying to compensate this with two DSL lines and therefore use Linksys/Cisco RV042 (v.1.1) routers.
The effect is sort of loss of the ability to establich an internet connection. We first of all thought that this might have to do with our dual WAN environment. We've gone through the procedure to name the usual suspects like:
ISP, Modem, Router, Cables, Setup (MTU etc.).
Our aim was to find a way to make it reproducible. Now we think that we have found a way to proof that this is a low-lying firmware problem. We've been using all available firmware for testing
RV042 firmware 1.3.12.6-tm,
RV042 firmware 1.3.12.19-tm and
RV042 firmware 1.3.13.02-tm
and we've used different routers of the same model to make sure it is not hardware dependent.
It cannot have to do with dual WAN setup, because we have physically disconnected the 2nd wan port and changed the setup so that the router should not have used this port.
We have been able to see this problem with a computer that is directly connected to one of the RV042's LAN ports, no other network hardware inbetween. When using another router (D-Link) the problem disappears.
This problem can be seen on linux systems using traceroute only. We've tried on windows as well, but do not see this. We think that's either due to
a) "The MS Windows tracert command uses ICMP echo request datagrams instead of UDP datagrams as probes." or
b) timing, because the tracert program seems to be slower by default than traceroute.
As you can see, slowing down traceroute by inserting "-z 500" makes the effect disappear.
So does it have to do with load as well?
There is no difference in the effect using traceroute with hostname or IP address. The effect is stable all the time.
To make it very clear we've made a small video which can be watched here:
http://www.youtube.com/watch?feature=player_detailpage&v=fbuJB8LqXeo
When calling traceroute three times after another, on the third call we ever see a delay/timeout. To us it looks as if this occurs every 8th time!
Does someone have an explanation for it? Any idea how to fix it? Can somebody check it on RV016 or RV082 routers?
Help appreciated!
Michael
Dear Forum.
Occasionally we experience problems with our internet connection. We are tracing this problem for quite some time now.
Due to the fact that we are little far away from the DSLAM we do not have the fastest DSL. We are trying to compensate this with two DSL lines and therefore use Linksys/Cisco RV042 (v.1.1) routers.
The effect is sort of loss of the ability to establich an internet connection. We first of all thought that this might have to do with our dual WAN environment. We've gone through the procedure to name the usual suspects like:
ISP, Modem, Router, Cables, Setup.
Our aim was to find a way to make it reproducible. Now we think that we have found a way to proof that this is a low-lying firmware problem. We've been using all available firmware for testing
RV042 firmware 1.3.12.6-tm,
RV042 firmware 1.3.12.19-tm and
RV042 firmware 1.3.13.02-tm
and we've used different routers of the same model to make sure it is not hardware dependent.
It cannot have to do with dual WAN setup, because we have physically disconnected the 2nd wan port and changed the setup so that the router should not have used this port.
We have been able to see this problem with a computer that is directly connected to one of the RV042's LAN ports, no other network hardware inbetween.
This problem can be seen on linux systems using traceroute only. We've tried on windows as well, but do not see this. We think that's either due to
a) "The MS Windows tracert command uses ICMP echo request datagrams instead of UDP datagrams as probes." or
b) timing, because the tracert program seems to be slower by default than traceroute.
As you can see, slowing down traceroute by inserting "-z 500" makes the effect disappear.
So does it have to do with load as well?
There is no difference in the effect using traceroute with hostname or IP address. The effect is stable all the time.
To make it very clear we've made a small video which can be watched here:
