Network loop being blamed on new equipment... help!

[DirectWireless]

I have a customer that I installed a Hotel Wifi network on their existing broadband connection ( a small local DSL company). I have a Linksys wired router that separates the Wifi from the Hotel’s LAN.

I used a IWE-1100A as the WDS master, inside the building, wired to the Internet, and the 2 Mikrotik routers each set as AP bridge with WDS enabled/static. On the more distant MT, I set up the IWE-1100A as a WDS interface, but did not bridge it. I also turned STP on both MT AP’s. One is using 2.9.12, the other is 2.9.8 (upgrade ran out on the CF card I used).

Now, the hotel complains their DSL connection drops off every few hours, and they have to reboot their DSL modem (a Netgear). It did the same thing when they had a Cisco PIX with DSL which was replaced when they initially had problems. The FIRST time they called their DSL provider, the girl on the phone said they had a bad tape drive lockup a RADIUS server. Every time after that, they were told nothing is wrong. They have had a very bad attitude at the provider all this time as well.

I have replaced switches, even went as far as disabling the farther AP(second hop), and it still locks up. The DSL company swears it is NOT their service, it is a network loop causing the lockups, which makes some sense in that rebooting the Netgear fixes it for a few hours. Nothing else stops working when the loop occurs, just the Internet, and it is fixed by rebooting the Netgear router.

[spire2z]

Don’t those consumer ADSL routers lock at the drop of a feather anyway?

An ant in Japan snezed and an ADSL router crashed in Outer Mongolia.

Why not make a Linux router with a Sangoma PCI adsl card. At least then if it does crash you can find out why?

[sten]

@DirectWireless

Now, the hotel complains their DSL connection drops off every few hours, and they have to reboot their DSL modem (a Netgear). It did the same thing when they had a Cisco PIX with DSL which was replaced when they initially had problems. The FIRST time they called their DSL provider, the girl on the phone said they had a bad tape drive lockup a RADIUS server. Every time after that, they were told nothing is wrong. They have had a very bad attitude at the provider all this time as well.

If you are not intimately familiar with how STP works then i suggest turning it off. DSL provider has tape drive -on- the actual RADIUS server? They do not have several RADIUS servers? Do they pipe their mail through this box too? Does it double as a Counter Strike game server on the weekends?

@DirectWireless

I have replaced switches, even went as far as disabling the farther AP(second hop), and it still locks up. The DSL company swears it is NOT their service, it is a network loop causing the lockups, which makes some sense in that rebooting the Netgear fixes it for a few hours. Nothing else stops working when the loop occurs, just the Internet, and it is fixed by rebooting the Netgear router.

What you describe does not sound like a loop, not even close.
If the switch ran STP ports would also close. All switch leds would usually flash continously. Your equipments ethernet ports will be flooded with unrelated packets. STP capable switches wont suddenly think it’s “OK” when you reboot the modem.

One thing i hope though, is that if the DSL modem is bridged, that the switch is not plugged straight into the back of it, but instead their firewall is. Some DSL modems do not support more than 20 or so mac addresses. Many support ~ 255 mac addresses.