Setup:
2x RB532A routing two LAN segments across 5GHz wireless link
Tx/Rx Signal is about -47/48 dBm, SnR 56 dB, CCQ is always at about 98-100%
Side A - Main office, Side B - Remote office, Both running RouterOS 3.2

Problem:
Clients on remote side of wireless link lose connectivity 3-4 times a day. All IPs on remote side are unpingable except for RB's ether1

Info:
I've been using the Dude from Side A to monitor the outages and discovered the the following:

The IP of ether1 on the B side is pingable without fail (24/7) from the A side of the wireless link. 3 to 4 times a day all other IPs on the LAN segment of side B die for anything from 2min to 2-3 hours. A reboot of the Side B RB fixes the problem after a minute or two. Have replaced the LAN switch on Side B as problem looked local as wireless link appeared up all the time.

Problem appears less frequently (3-4 times / day) since upgrade to RouterOS 3.x stable (went through all 2.9 releases hoping it would disappear) but is incredibly annoying to users as some VoIP phones are located at Side B.

I have enabled the Watchdog on Side B RB to ping another RB that provides local Wireless at that site, but I would prefer to have the problem disappear completely and have the connection as rock-solid as the wireless appears to be.

Please any suggestions.
--
Joe

To follow up:
Upgraded both sides to 3.4 last night.
Problem worse! And reboot now doesn't fix it!

Via winbox from Side A to RB on Side B shows that the ether1 interface has "no link" in the status bar but under the Status tab we're connected at 100Mbps Full duplex.

Am going to manually switch cable over to ether2 and see what happens.

Greetings!

Are you using the dynamic WDS bridge setup (wds-mode=dynamic) or EOIP? If it is WDS, insure you have assigned the wds-default-bridge in your "/interface wireless" settings on the AP side. The docs warn that if this is not set, the bridge may not reconnect if the signal is temporarily lost for any reason.

I have flogged my system thoroughly. I kill power on each box to test them. Not a shutdown, I kill the power. And no matter what I do to it, the bridges always reconnect now.

Thanks for your reply.

Unfortunately I'm not using WDS and routing everything - no bridges.

Have switched over to using ether2 for traffic and injecting PoE on ether1 as a test.

Futher information gathered:
Sitting at LAN on Side B sees the ether port on the RB slowly die after an hour or two with ping responses getting worse and worse to it.
Power cycling the unit brings it back to life. This occurs using ether1 or ether2.

Am I wasting my time troubleshooting a board that's faulty or is there another possibility why the board is behaving this way?

Did you ever find an answer to this ?
I am having the exact same problem, the wireless link stays up but no data passes after a while, a reboot of the B end fixes it again for a variable period of time.

Regards
Paul

Hi,

Unfortunately, no. Ended up replacing both ends with RB433 boards, leaving radios and antennae as is. Link has been 100% rock-solid ever since.

Funnily enough I have been using the RB532 boards as access points in my house for over 6 months now and they too seem to be working fine!