Re: [WISPA] Odd Speed Results

2014-04-18 Thread Marlon Schafer (509.982.2181)
Look at a difference in rssi levels at each end. I've seen some of the radios (xr2 and xr5 cards) go weak either on the tx or rx side. That's a hard one to catch. Also if it's only 2.4 miles you might have to much power going to them and picking up some multipath. Try dropping the power

Re: [WISPA] Odd Speed Results

2014-04-18 Thread Robert
Ubiquiti web test = udp and Mikrotik = udp or tcp so which are you using on the Mikrotik side udp or tcp?IF you are using tcp on the Mikrotik test, most likely you have a lot of packet loss and the tcp is spending a lot of time in retransmission. If it's udp vs. udp then start checking your

Re: [WISPA] Odd Speed Results

2014-04-08 Thread Scott Reed
At daybreak. Already have a climber scheduled in case it is at the top (265') On 4/7/2014 10:05 PM, Mike Lyon wrote: Check the cable and hard set the interfaces if you can. Cheers, Mike On Apr 7, 2014, at 6:52 PM, Scott Reed sr...@nwwnet.net wrote: Oops, look at the right stuff.

Re: [WISPA] Odd Speed Results

2014-04-08 Thread Scott Reed
Web on UBNT, WinBox on MT On 4/7/2014 11:36 PM, Robert wrote: I didn't think the shell based speedtest would give those numbers without limiting out the cpu. Were you using the web based speed test? On 04/07/2014 06:37 PM, Scott Reed wrote: I have a link using 2 PowerBridges that is about

Re: [WISPA] Odd Speed Results

2014-04-08 Thread Robert
Web on UBNT is a UDP test.. Apples and oranges As the shell based tcp test on UBNT hits processor limits very fast, it's better switching the MT test to UDP to get a better comparison. Best, Robert On 04/08/2014 03:07 AM, Scott Reed wrote: Web on UBNT, WinBox on MT On

Re: [WISPA] Odd Speed Results

2014-04-07 Thread Mike Lyon
How do the interfaces look? Are they negotiating to 100/full on all sides? Any interface errors accruing on any of the devices? On Apr 7, 2014, at 6:38 PM, Scott Reed sr...@nwwnet.net wrote: I have a link using 2 PowerBridges that is about 2.4 miles. I can get 56M TCP across the link using

Re: [WISPA] Odd Speed Results

2014-04-07 Thread Scott Reed
All are 100/full No errors On 4/7/2014 9:42 PM, Mike Lyon wrote: How do the interfaces look? Are they negotiating to 100/full on all sides? Any interface errors accruing on any of the devices? On Apr 7, 2014, at 6:38 PM, Scott Reed sr...@nwwnet.net wrote: I have a link using 2

Re: [WISPA] Odd Speed Results

2014-04-07 Thread Scott Reed
Oops, look at the right stuff. 760,000,000 tx errors in 79 days on the wire from the RouterMaxx to the PowerBridge. Guess I know where the problem may be. On 4/7/2014 9:49 PM, Scott Reed wrote: All are 100/full No errors On 4/7/2014 9:42 PM, Mike Lyon wrote: How do the interfaces look? Are

Re: [WISPA] Odd Speed Results

2014-04-07 Thread Mike Lyon
Check the cable and hard set the interfaces if you can. Cheers, Mike On Apr 7, 2014, at 6:52 PM, Scott Reed sr...@nwwnet.net wrote: Oops, look at the right stuff. 760,000,000 tx errors in 79 days on the wire from the RouterMaxx to the PowerBridge. Guess I know where the problem may be.

Re: [WISPA] Odd Speed Results

2014-04-07 Thread Robert
I didn't think the shell based speedtest would give those numbers without limiting out the cpu. Were you using the web based speed test? On 04/07/2014 06:37 PM, Scott Reed wrote: I have a link using 2 PowerBridges that is about 2.4 miles. I can get 56M TCP across the link using the UBNT

Re: [WISPA] Odd Speed Results

2014-04-07 Thread Robert
What version firmware on the RouterMaxx? On 04/07/2014 06:52 PM, Scott Reed wrote: Oops, look at the right stuff. 760,000,000 tx errors in 79 days on the wire from the RouterMaxx to the PowerBridge. Guess I know where the problem may be. On 4/7/2014 9:49 PM, Scott Reed wrote: All are