I have a Rocket M5 XW on firmware 3.20.3.1 that seems to 'erase' itself when I save a change and reboot.
I've had this Rocket for about 18 months in testing, but recently deployed it up on a tower. I belive the upgrade to 3.20.3.1 was rather recent for this device. I didn't see this happen until of course it went up.
I was able to recover it by doing a factory flash on the near end using the reset button. I did the flash sequences, configured the node, saved the settings, and it was working fine.
For the second time now, I would make a slight change, save it, and reboot as needed, and the device disappears never to come back. I have two other Rocket M5 XW that have been in service on firmware 3.20.3.0, never had this happen.
My next thing to try will be to flash 3.20.3.0 instead of 3.20.3.1, and see what happens. I guess I would hope to see if I can recreate the issue after stepping back.
I guess my surprise is that I am able to flash and save settings initially, but at some point in the future that seems to bomb out.
Anyone else been in this boat?
Matt
You are here
Rocket M5 XW node reboot issue
Sun, 12/20/2020 - 18:13
#1
Rocket M5 XW node reboot issue
Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer
I had a similar issue with a Nanostation. Turned out to be an issue with the reset being asserted for some reason, resulting in full clears of the configuration as well as the mentioned AWOL behavior. I was able to put a stronger pullup/pulldown resistor on the reset line to fix it. I was able to scrounge up the necessary information on the reset circuit online to effect this repair.
Turns out nothing has really been erased, but its obvious for some reason the power up is not working properly.
We did test the shielded cable from the rocket to the power injector 'POE' port when the unit was installed, it was OK (supposedly, right?)
Anyways, so just removing power to the injector and re-adding it did not resolve anything. Also, disconnecting the shielded cable from the 'POE' port and re-inserting it did not help either.
I also tried a different power injector, which DID work the first time, but subsequent reboots did NOT.
Then, I tried all sorts of fancy machinations with power, cables, etc.
What DOES work is if I do this:
Remove POE and LAN cables. Remove power from injector. Wait about 15 seconds. Re-add power to injector. Plug in 'POE' cable, count to 5. Re-add 'LAN' cable. Then, it works.
We have a Mikrotk hAP on the LAN port. The piece of CAT from 'LAN' to port 5 is NOT shielded. Could this be an issue? I would be surprised, becuase I can think of anoter install we have, where the 'LAN' to port 5 is NOT shielded, and the 'POE' to Rocket IS, and never had an issue.
Also, the Mikrotik IS being powered by 12 volts, rather than 24 volts. My only though is that, for some reason, the Mikrotik has some kind of voltage out on port 5 (which is NOT enabled, btw), and the power injector is disabling it's own power out. If this were possible, then the 12v from the Mikrotik will not obviously be able to send enough power down 100ft of CAT5.
So, for now, we have a 'method' to get the Rocket powered up, and a few things to try next time:
Change the 'LAN' to Mikrotik cable to shielded
Use 24v adapter on Mikrotik instead of 12v
Honestly, other than trying these two changes, I have NO idea what is up, hah!
I seem to recall that the hAP lite will momentarily assert power on the POE-out port on power reset/restart and then turn it off late per switch setting after the POST process. If so, I wonder if that was the culprit. Your inserting a dumb switch in between was a good idea and effectively filtered out any POE power coming thru that connect. Thanks for that reminder.
I've have other strange things happen in the past with various Ubiquiti nodes from time-to-time and I wonder if these types of ghost resets were the root cause. Oh well, it's all experimental, right? Perhaps it almost makes sense to have cables with defeated reset pins for when you don't need remote reset.
- Don - AA7AU