BH links bouncing at night now....

For the past few weeks, all my BH links (about 10) have been bouncing every 10 mins… form the hours 11:30ish pm to 6:30ish am. I see no errors in any of them, M or S sides. they just bounce every 10 mins- up for 10, down for 10. They do mostly seem to be south of our main building (up to 18 mi south). there isn’t a storm on those nights either but they do seem to be worse on nights that fluctuate more from the daytime temp. Could this be dew or condensation? Once daybreak is up for a while, they are stable all day. A few links bounce a few times, others do all night every 10 mins.

Any ideas?

Thanks a million!

How long are the links?

When you say they are bouncing, is that something a possible faulty monitoring system is alerting you to, or is it true to be down?

I use Nagios to monitor all my devices, we have many many towers with anywhere from 6 to 16 AP’s per tower and of course the backhauls.

What I noticed about 6 months ago, was that whenever my Monitoring server was under heavy load or being backed up that it would generate false alarms. When pinging, telneting or logging in via my web browser all was ok, and the log files reflected that no reboots were noted either.

On the 10 Backhauls, are they all terminated to the same piece of equipment? i.e. your masters.


Thanks for the help,

For the first question, the links are from 5 to 12 mi.

The monitoring system was first on my suspicion. I will more closely watch it as I troubleshoot this. We have it set up where (ip monitor) will report a link down after failing 3 consecutive 300s(the monitoring ping), then two 100 secs apart (after a failed ping). As for the termination, we have three main points of connections- then the rest spider form them. those points adhere to the all masters or all slaves on them and have CMMs. outside these main points, anything 2 hops away is timed via pipesync. Aps on those over towers are likely timed via timing port on the back-haul (unless more APs are there, then a pipesync is also used to time the APs. The only deviation at this time is one back-haul that is timed from a back-haul (one hope only). However, this has always been this way and the bouncing is not on that link either (its actually on the other side of the network).

As I think of it, I have found a few links to be running software instead of hardware and I’m fixing on ironing those out as well. Since I did not put those links up, I sometimes worry if maybe hardware wouldn’t work on those links. it’s available, but not in use. If it works in software (link wise) shouldn’t it work in hardware? Other than that, once I manage to get this network to par, I will be upgrading firmware as I am on 7.2.9 right now and I imagine some improvments should be done in new versions by now.

"""" New information""""""

I have checked now when the pings seem down, and they really are down. Actually, for the night form 10"30 pm to 7 am, the ping times are about 5000ms when I can get a response. Now I’m wondering if my radius is not propagating upload limits and people are letting their p2p going. Seems like a possibility, and I will look into that while I wait for some ideas :).
– LP


A couple of ideas, it does seem very odd that it only happens at night. I don’t think it has anything to do with temperature or climate.

Has power been ruled out? Is there possibly someone else colocated performing maintenance in nightly Windows? This could be ruled out if it happens every night, I know I hate my Windows and would not do them 7 days a week.

I would change eveything to Hardware Scheduling, you will have to go there very soon anyway, because of the path Motorola is taking with all their new equipment, and update the firmware to the latest as well.

We just finished, or I should say are almost finished migrating our 3K customers from Software to Hardware Scheduling. During this time frame we had overlapping sites casuing self interference between software and hardware until we could make all the equipment changes needed.

Wish we would have been on Hardware all along, there are definite management benefits, the AP Autoupdate Feature Actually works and works well in CNUT.

The other thing hardware has allowed us to do, is extend our network into areas where other competition was, hardware schedulign just was chewing us up whilst in software mode.

The other thing I would do, is change all your radios to temporarily force 10BT Full Duplex, this will rule out any Auto Negotiation Issues that may be arising because of interference in in the cable.

Grounding? Might sound silly, but at one of our sites that also houses Cingular Wireless Equipment, they changed something about 7 weeks ago, I had flapping AP’s and backhauls when they did that. Finally after beating our heads against the wall, we decided to disconect the shared ground to see if there was a difference, and that fixed all the problems.

We ran our own isolated ground and all was back to normal.

And lastly, outside interference, have you tried different freqs?

Hope this helps.


I assume this is new? New, random issues are almost always interference.

I would start looking for a new RF source in the vicinity of where your BH links come together.

Are you certain that it’s multiple links, or could it be one link common to the others?

Thanks for your suggestions!

These are definitely different links, well, for the most part. There are some competitors in the area and unfortunately, the one that is starting to use Canopy will absolutely not work in coexistence whatsoever. Another strange fact, they all are bouncing for about 5 minutes on the dot. Last night they did it again, and all the times were the same. With monitoring, it is really hard to determine sometimes what really went down. I will recheck the logs and check all of them. I know they are not loosing power as their uptime is continued on the status page of the APs and BHs. I will check all grounds as well. I hope that the problem is not similar to the ranging diff I’m getting on some APs.

Perhaps a reboot may be in order, then, so will updating. I will begin the switch to the 7.3 version next week- then it will be up to 8.5.1. I’ll check all the grounds and logs. As for interference, I am constantly wondering as we all probably do, but why just at night ya know? As for the software mixing … some of these backhauls don’t have a hardware option. Is that update-fixable or just too old and will have to be replaced?

Thanks again and i will post what I find.


Read this thread on uncooperative WISPs … hlight=fcc

Set em up with the letters, then knock em down with your letter to the FCC. CC the other WISP on your letter to the FCC and watch em come running. If you need help drafing your letter to the FCC let me know.

You were there first, you win. The hard part is getting the FCC to care.