so a few updates on this since the other jay first posted it - 

* we've seen this before, sometimes with ubnt gear.  i guess the proper thing 
to do is wireshark the connection but since the towers are many miles away and 
it was cold and raining, we just started stumbling through troubleshooting 
steps late last night

* at times a ubnt radio (haven't seen it with any other radio) will "flood" the 
network causing lost packets - i assume due to too much traffic or some other 
cause.  there is one nanobridge feeding two sites that are within 8 degrees of 
each other (one 9 miles, one 12 miles).  each site started with only 900 mhz 
canopy (three sectors a piece per site).  we have added one 2.4 ubnt omni 
within the last year to one of the sites.
we added 2.4 epmp with a 120 sector to the other site within the last year.

* as stated last night, we could disable either site "a" or site "b" and have 
no packet loss.  we could not run them both without packet loss.
when going to bed, we left the busiest site up (with the most business 
customers) and the smaller site was offline.  these are pretty busy sites, 
probably 70 customers between both towers.  the split was probably 40 on one 
tower and 30 on the other.  again, 3 meg service (mostly)
so although they'd fed by ONE nanobridge (feeding two nanobridges at each 
tower), it was in line to move to ac gear.

*  the theory was we were going to have to replace the ac radio that is on the 
feeding tower (the master side).  In hind sight, I think doing that would not 
have resolved the problem.

*  they are on similar subnets  (this segment is bridged, not routed)

*  when i got back on this today i disabled the 2.4 omni on site (b) and left 
sites a and b connected.  there was no packet loss.
this led me to believe the problem lived on the 2.4 omni on site (b).

*  there are 11 customers on site b's omni.  I began changing each customers 
ssid to "testsite" and dropped them off one by one until i identified the 
problem customer.  i then changed all the other customers to "testsite" and 
moved the main omni ssid to testsite, leaving the problem customer offline.  

no more packet loss.

This did start after a line of storms moved through yesterday.  I assume she 
took a close strike but the radio did not die.



  ----- Original Message ----- 
  From: Jay Weekley 
  To: Ubiquiti Users Group ; af@afmug.com 
  Sent: Thursday, March 2, 2017 8:22 AM
  Subject: [Ubnt_users] Odd Nanobridge problem


  We have a 5.7 Nanobridge acting as an AP that serves two tower sites 
  that are close together.  After a storm yesterday we started getting a 
  lot of packet loss when pinging the AP.   We can disconnect either of 
  the SMs and the remaining tower works fine and there is no packet loss 
  to the AP but when we turn both sites back on the packet loss returns.  
  There is also a Netonix switch at the tower that has the 5.7 Nanobridge 
  acting as the AP and it's cable diagnostics tool shows a problem on pins 
  5, 6, 7 and 8 though the AP is powered up. Both these sites are slated 
  for upgrades to Ubiquiti AC gear but the logistics will make it 
  difficult this week.
  _______________________________________________
  Ubnt_users mailing list
  ubnt_us...@wispa.org
  http://lists.wispa.org/mailman/listinfo/ubnt_users

Reply via email to