Did the APs lose power when it went offline? Does the problem happen only on particular APs?
How many E410 are connected to this switch? The EX2028-P has the max PoE budget of 400W. I doubt that the switch is low on PoE power in which it will start pruning power on the AP. The AP is classified for class 4 which can draw max of 30W, but I have not seen they are using anywhere near 30W. Other possible cause for the AP to lose power is if the AP at some point draws more power than the max allowed for its class. The switch would turn off the PoE on the port momentarily when this happens. A workaround is to set the port’s PoE mode to hybrid. Last possible cause is something wrong with the AP.
Next time when the symptom occurs, SSH connect to the switch and execute the following commands:
show system info
show power inline
show power detail
show power inline measurement
show logging
For me it looks like this.
The switch reboots and then one or two devices do not get up.
The port shows a voltage of 2.4 W, you cannot ping devices.
Just Toggle PoE and the hotspot will start up normally.
I don’t know if this problem occurs on 5.0.1, on older versions it was noticeable.
Thanks Pshemo for the description.
What’s the cnpilot model? What’s power consumption by other ‘working’ APs?
When these 2 devices are working after toggling switch port’s PoE, what’s their power usage?
On the switch can you see the AP’s LLDP info (CLI command ‘show lldp ne detail’)?
I will try to reproduce the symptom in the engineering lab. Can you provide me the following details?
switch’s software version
cnpilot model & software version
number of cnpilots connected
Is the symptom intermittently or does it happen every time switch reboots?
I have the following setup in my lab. I have not seen similar symptom so far.
EX2028-P on 4.4-r4
12 E410s connected to EX2028-P. No Wifi clients.
All APs came up, each consumed about 4.5W
Powered cycle the switch
All 12 APs came up and consumed ~4.5W after some time.
I didn’t run the ping test or cnmaestro online test as these APs are not onboarded. However, mac addressess of all APs are learned on the switch. LLDPs are learned for all APs.
It doesn’t seem to be the case that the APs draw more power than its class, so hybrid PoE mode isn’t useful here.
Please raise a support ticket so we can assist you further.
Hi Toban, how many XV2-22 are connected to the same switch? After switch reboot, did some XV2 fail to connect online? Did you toggle the PoE power to recover the APs? Does it happen every time?
Can you please share the output of CLI command ‘show tech’?
No more than 12 are connected to any given switch. In my case, a reboot of the AP will usually bring them online. In some cases, I change the DNS address then back to the original and then they reestablish communication without rebooting. I have a couple that I can SSH into and can’t HTTPS into. I can get them back online, then some will drop back off after a couple of hours. It is as if a buffer is getting overrun.
Suddenly pull the plug out of the socket and immediately plug it back in, something like a power surge.
This is when it most often happens, but recently after upgrading from 4.6 to 5.0
There are some HTTPS and HTTP problems if the user tries to access the UI of AP in 6.6-r8. We are addressing HTTPS and HTTP problems in an upcoming release.
Hybrid port is not the solution (the problem still occurs), we upgraded the sw from 4.4 to 5.0.1 and after sw rebooted, 1 of 3 AP didnt show up on9 in cnm, (can’t ping the AP) (had to disable POE at the port and enable it back) . Attached is the sw tech file after the reboot and the 1 AP already come up on9 in cnm. device_response.txt (94.9 KB)
Reporting to support seem like no point since AP / SW tech support files is not usefull to finding the problem. can refer # 352380 , we stop the follow up sincewe need to invite support and we cant do that because the account is not control by us and cannot get approval.
I can say one thing. On version 5.0.1, there are no problems with starting cnPilot e410 after a power loss.
The cause was versions 4.4 to 4.8.
The symptom that several switches were not synchronized with the controller settings after restart also disappeared. Now there is no such thing.
Thanks Pshemo for the update. It would be interesting if other users (Toban, lady.m3lody) have similar result with 5.0.1.
Regarding “The symptom that several switches were not synchronized with the controller settings after restart also disappeared”, do you mean cnmaestro shows switch not in sync after switch reboot? Do you see this behavior on more than one switch?
As replied, we did upgrade to 5.01 but problem still happen.
its happen to about all the sw in the network if there a reboot on the sw (5 - 7 switches per sites).
The sw is in sync after upgrade fw. or after reboot
The lack of synchronization only occurred in my case on 1028 with automatic device recognition set. After restarting, older versions of the software always showed not in sync. I had to manually force synchronization.
There were only cameras on these switches.
Thanks lady.m3lody for the update. My understanding is that you are still seeing 1-2 APs offline after switch running 5.0.1 reboots. The symptom is present on several switches.
)
