SMART-RF malfunctioning

[winger]

Hello, Extreme community! 

We have a problem with the SMART-RF functionality in our deployment. In a nutshell, something is causing "Select Shutdown hide" in an RF-Domain that has its own dedicated SMART-RF Policy with "no select-shutdown" configured. We have multiple RF-Domains on our controller covering one region of the globe. 

We are running: 7.9.0.1-007R on our NX-7530-100R0-WR controller and a mix of different firmware on the APs. They are like this: 
ap7532 > 7.7.1.6-006R
ap7632 > 7.7.1.6-006R
ap305c > 7.9.0.1-007R
ap305cx > 7.9.0.1-007R
ap410c > 7.9.0.1-007R
ap410-1 > 7.9.0.1-007R

I have checked the local RF-Domain manager config for any possible config mismatch, and I didn't find anything different from what there is on the central controller. There is also no reason/indication from SMART-RF to hide the R1.

With that particular RF-Domain we also notice something funny. From time to time when show wireless ap on XXX is done, the result is empty. And after some seconds it's filled with data again. Anyone faced similar behavior? Here are the SMART-RF Policy details: 
sensitivity custom
assignable-power 5GHz max 18
assignable-power 5GHz min 14
assignable-power 2.4GHz max 18
assignable-power 2.4GHz min 6
channel-list 5GHz 36,44,48,128,40,149,153,157,161,165
smart-ocs-monitoring sample-count 5GHz 10
smart-ocs-monitoring sample-count 2.4GHz 15
no smart-ocs-monitoring power-save-aware 5GHz
no smart-ocs-monitoring power-save-aware 2.4GHz
no select-shutdown
no smart-sensor
no smart-sensor auto-trigger
smart-sensor band smart-band-5GHz
no smart-sensor tri-radio-only
no coverage-hole-recovery

[winger]

Hello Daren,

thank you for pointing in the direction of the FW swap. We are considering it. I would really like to drop all of the non-universal APs: 7xxx series and stay with 310's/305's/410/460's. Unfortunately, this is not an option for us right now. I was looking at the SMART history from the controller and I got this: 
% Error processing 'show smart-rf history on XXXX':
Traceback (most recent call last):
File "/bld/n03/workspace/ADVAP_n_CTRL/wing/obj/nx7-bare_metal/src/sys/cfgd/cli/show.py", line 130, in _show_read
File "/bld/n03/workspace/ADVAP_n_CTRL/wing/obj/nx7-bare_metal/src/sys/cfgd/wireless/smartcli.py", line 2872, in show_smart_history
File "/bld/n03/workspace/ADVAP_n_CTRL/wing/obj/nx7-bare_metal/src/sys/cfgd/core/cluster_db.py", line 324, in _generate_remote_xpath
Exception: Timed out waiting for remote device 1C.A2.2E.C0 xpath=wing-stats/rf_domain/XXXX/smart-rf/history

is it giving us any clue on where the main problem is?

[Daren_Ellis]

It sounds like the RFDM (rf-domain manager) is not responding.
Do a "show mint neighbors" and a "show global domain manager" and ensure that the device reporting as the RFDM has mint id 1C.A2.2E.C0 and is function properly on the network.

I would try a quick reboot of that device and see if that helps.

[Daren_Ellis]

Morning winger,

If you are not trying to take advantage of Nova UI i would suggest upgrading deployment to 7.7.1.10.
If you are trying to get away from the flash GUI please update deployment to 7.9.5, the AP76xx APs will update to 7.7.1.10 as they are not capable of 7.9.x code.