ePTP keep alive stuck problem is still out there

Supposedly fixed under 4.6.1 but we recently turned up a couple of mini pops and are using 2.4 ePTP links for backhauls due to trees. 20 Mhz channels with some noise spike but SNR is typically 29-31 and MCS levels are 13-15.

Both links are dropping their connection reporting KEEP ALIVE DETECTED STUCK, DISCONNECTIONG on the masters.

As noted in previous instances, it does NOT do this if in TDD PTP, but we need the throughput of ePTP

We are currently running 4.6.1and uploading 4.7 RC14 Beta and will check to see if that helps any.

There is also an RC15 beta posted in the beta group, and for us, it’s LOT faster. RC14 had a throughput issue, fixed in RC15 and above.

2 Likes

try dropping out of eptp and using tdd mode but limit which radio can associate with each ptp AP.
We are running 4.6.1 on a 2.4ghz link with no problems

DITTO. We are using all our Force300 backhauls in TDD 75%/25% mode. For us, ePTP mode on the 1000/2000s was MAGIC, but on 3000/300s it’s been constantly problematic. We’ve gone to TDD 75%/25% and had no issues. Besides… with all the issues we’ve has with ePTP, the ePTP mode seems to be much slower than expected, so TDD for us works just as fast anyway.

It appears its more problematic on 2.4 than 5 Ghz, not sure why that would be. We really needed the ePTP for the latency . Our (stupid) motto is that we will go where others won\t. We are attempting a micropop/spider web scenario way, way out in the boonies. It’s a 9 mile 2.4 back haul clipping some trees. Its already a double back haul and this will make it a triple so trying to keep latency somewhere useful, though since we are competing with satellite, its a pretty low bar to overcome, but need to be able to support video conferencing and VOIP (and gaming).

We recently swapped out one of the F300 BH links in TDD PTP with a 5 Ghz Air Fiber unit, which lowered the latency (and jitter) dramatically.

Right now using ePMP 1000 in ePTP gets us usable bandwidth out to the woods, but if demand gets very high we will probably try out an Air Fiber 2.4 back haul. We considered a 900 Mhz PTP but its pricey and the overall bandwidth is not that awesome.

On the plus side, we upgraded to the beta 4.7 RC14 on the ePMP 1000 2.4 units and it appears to have corrected the issue. We now have two such links running on the beta and neither have blipped since. 4.6.2 may have also corrected it, but never loaded it as we were waiting for 4.7 to be officially released.

if low latency is your goal, then you really need to look at the ptp5xx or 6xx lines. If you need low latency and high throughput then ptp8xx is your beast.

2 Likes

Unfortunately, that’s what we’ve done too. We still have several F300 backhaul links where our requirements are low, and in TDD they work ok for us, but they all get swapped for AF5xHD sooner or later.

2 Likes

We had replaced every single ePMP PTP except 1 with Ubiquiti 60, 11, or 5Ghz AF. The one we had left was PTP550 that I was going to replace with Ubiquiti because all 3 (the pair and the spare) was almost daily either rebooting or shutting down the link for DFS. I’m desperately trying to stick with one solution so I thought I would give AX a shot… gotta say, after an initial config issue, it’s at least as reliable (so far anyway) as the 5xHD’s have been.

Kind of gives me hope that maybe ePMP will redeem themselves with AX

1 Like

We want to try the F400 for backhauls, if we could ever get any. 4.7 RC14 has not correctedthe problem with the ePTP and KEEP ALIVE STUCK issue. We have a problem link on a small backhaul to a mini setup and it is consistently dropping out, sometimes in hours, sometimes a couple of days. The Master then comes back reporting is has no registered subscriber modules yet it is lying as the link is up and passing traffic… On several occasions when it happens again while in that state, it will then report it shows 65000+ registered subscriber modules and while it shows up, traffic does not pass and requires a reboot to recover, and the entire process starts over. We opened a ticket and uploaded the maintenance file captured while in this state.

We have another 2.4 ePTP backhaul link in a much more rural area that seems to be running okay, so maybe some flakey hardware? We will try swapping the master this weekend