Strange ping problem, RB112 + SR2 + CM9

RouterOS Wireless Networking
1

Hy all,
this is my setup:
RB112 + SR2 + CM9 onboard, SR2 for hotspot and CM9 for backhaul.
When i ping flood the ip of RB112 via wireless (bridging beetween ethernet, SR2 and CM9), either to SR2 or CM9, the ping stays under 20ms, and every 3 second (terribly constant) it goes up (for a single ping) to 250-500ms.
This is very annoying, becouse the TCP throughput goes down every the same 3 seconds, from 5Mbps to 1,8-2,2Mbps (but no timeout or disconnection), and then it return to 5Mbps.
I tryed with a Linksys WRT54G/GL as client, and with Ubiquiti SRC PCMCIA client, but it’s the same thing. SR2 and CM9 are set on 2,4GHz band. SR2 is “only-b”, becouse, in “g mode” or “b/g” the ping goes terribly high (and, this is strange a lot for me, but it’s another problem…).
I tryed to disable one card, and then the other, but same thing happens. I tryed to disable radio calibration, polling, ssid broadcast, default authenticate, put fixed data-rates, ack-timeout from dynamic to static, compression…but no difference from the initial setup.
Is this a RouterOS bug (as i read on other post) or other?

Thanks.

2

Just tried to lower the power out of the SR2 to 20dBm: no result, same things.
When I made the ping flood, CPU stays beetween 2% and 6%.

Additional info: rb112 with a omni antenna 12dBi, client with a rootenna 14dBi, signal received by the client -72/-76dBm, signal received by RB112 -81/-86dBm, bitrate 11Mbps rock-solid, CCQ (on RB112 side) 98/100%.

This is an output from a linux machine:

--- 172.21.1.1 ping statistics ---
9800 packets transmitted, 9799 received, 0% packet loss, time 57681ms
rtt min/avg/max/mdev = 3.102/16.585/477.926/51.086 ms, pipe 39, ipg/ewma 5.886/3.580 ms

And this is an output from look@lan:

3

I have this issue without flood ping. :frowning:

4

Update:
today the problem is disappared, but I touch nothing, neither on ap or client side!
Is MT weather-sensitive? :wink:

5

what kind of weather you have in your area? :wink:

6

‘floods’ :slight_smile:

7

:smiley: :smiley: :smiley:
Yes, we had “floods”, but many years ago! :wink:
Today, for the second day, the flood seems to be stable. The bug i illustrated here is no more visible. I’ll update this thread when this annoying “bug” will be again here, possibly with more details.
Is possible that this “bug” was produced by a small quantity of fog?

8

normis... !!!!
i had floods by mud almost 6 moths ago.. :cry: :cry:
http://www.flickr.com/photos/jrki/242355444/
that's 400km from us.

regards
Hasbullah.com