I’m running into some issues with setting up a Mikrotik RB112 Point to Point link.
I have 2 RB112 units each with a CM9 card running over POE. 1 is set to Ap-Bridge mode with WDS enabled (we’ll call it the host) and 1 is set to Station-Wds (we’ll call it the client). I’m using 2.4Ghz-G-Turbo with NStream and Polling.
I got the WDS link going and everything seems to work fine, the trouble is anytime I have more than about 5MB (this is a more or less number) going upstream over the client it reboots.
For example if I do a bandwidth test from the host RB112 to the client RB112 and don’t put any restrictions and do a SEND (so host to client) it seems to get up to about 8MB or so of throughput, which is great. Now if I stop it and switch to RECEIVE (so client to host) it starts up then after about 5 seconds it reboots (the client that is). I’ve tried this on the client as well using badwidth tests and it does the same on a SEND from it. I’ve tried file transfers and they die as well. I’ve tried adding a queue to the client to not allow anything over 5MB and this seems to work (most of the time).
So it seems if I limit it to 5MB upload over the client things seem OK, for the most part-though there are still some drops now and again, but over that and it dies.
Anyone have any ideas? Could this be a frame issue, should I lower frame size or something?
Sorry, but using a RB112 for a PtP link using NStreme is definitely too much.
This board is designed as a CPE device, and the CPU will choke on NStreme quite quickly…
Actually, we tested them, and they will run fine with Nstreme, just not very fast - say 10Mbps on a P2P link.
turn off Turbo mode, that’s not necessary or a good idea - and uses up still more CPU power. Also in 2.4GHz you are likely to suffer interference in the 40MHz band. Use 20 or 10MHz channels.
So I did ditch the NStream for this link - it’s a very short link and I wasn’t seeing any performance increase.
I talked to the guys over at Wisp-Router and they said that in many cases the RB112’s just use too much CPU (like you all mentioned) and give a kernel panic causing them to reboot.
I had my client RB112 set to Station-WDS, they suggested setting them both to AP-Bridge. I did this and eventually set them just to Bridge mode (both of them) with WDS enabled and this solved the reboot issue.
I can view the RB112 host and do a scan and see that the client totally drops it’s signal, ie. it stops broadcasting and then after a minute or two will come back.
If I leave the host in scanning mode, the client remains there, but if I make the host again try to connect it does for like 10 seconds then drops the connection and I do a scan and see that the client is again not broadcasting.
I think I’ve got my two radios overpowering each other. ie. I’ve got too much signal for too short a distance - it’s only like 1 block (probably less than that) and a straight shot.
I’ve been trying to reduce the tx power and that’s kind of helped, though my SNR has definitely dropped, but even with that it still seems to drop sometimes.
