Clients only connecting 11a/11g

[akcampbell]

We currently have two 9800 WLCs serving our schools. Their configs are virtually identical. When clients connect to one specific SSID on 9800 A, they'll connect at expected speeds using 11ac or 11n protocols. However, when clients connect to the same SSID on 9800 B, they only connect using 11a or 11g protocols. RF settings appear to be the same across both controllers. Any ideas?

[Mark Elsen]

 - Have a checkup of the  9800 B controller's configuration with the CLI command show tech wireless
   and feed the output from that into : Wireless Config Analyzer
                     Adding : do the same on 9800  A and look for differences.

  M.

[akcampbell]

Will do! Thanks for the suggestion.

[MHM Cisco World]

Can you share RF config 

MHM

[akcampbell]

We're still new to the 9800s, so I'm not entirely sure how to tell which RF profile is being used for that specific SSID.

[MHM Cisco World]

Configuration > wireless > access points

This will list all AP's

Beside AP name there is blue icon click it 

You will see the rf tag of AP

Compare this for two wlc and check different

MHM

[akcampbell]

Ahh right. 2.4 and 5 are both using global configs, and the actually RF tag for the site says global config as well.

[Leo Laohoo]

What model of APs in 9800A and 9800B?

What firmware are 9800A and 9800B on?

[akcampbell]

3802is for the most part, with a few exceptions.

IOS version 17.9.4.27

[Leo Laohoo]

---

@akcampbell wrote:
3802is for the most part, with a few exceptions.
IOS version 17.9.4.27

---

A perfect combination (not good). 

Reboot all the 3800 and see if this problem goes away.  

If the problem goes away after the reboot, consider moving away from 17.9.4/17.9.4a or 17.9.5.  We are seeing a significant amount of really "bizarre" bugs.  

And as a "bonus" for using 3800 (which is another bucket of bugs due to a hardware design bug in itself), regular (or daily) reboot of the 3800 (regardless of firmware version of the WLC used) will also make significant improvements.

[Rich R]

> We are seeing a significant amount of really "bizarre" bugs.
We've seen one a few times now on AireOS 8.10.190.0 when after a site has switched between WLCs (N+1) and then back to primary WLC all the APs start silently dropping *lots* of incoming packets (sometimes up to 50% even at fairly low traffic levels) so CAPWAP starts flapping.  AP reboot fixes it.  TAC (after months of captures, debugging and log collecting) still have no explanation.  They were only able to establish that the packets are being dropped at a very low level in the AP (possibly at ethernet driver) but do not know why...  Due to another known bug (which will not be fixed) the captures and debugs were not able to confirm why the packets get dropped.  Since AireOS is EOL they are not doing any more about it.  We now know that when this happens we need to reboot every AP on site!

We're relatively stable on 17.9.4 + SMU's and APSP6 at the moment with C91xx, 1832 and 1562 APs and starting to look at 17.9.5 but waiting for TAC to explain a WLC crash on 17.9.5 seen in lab.

[Leo Laohoo]

---

@Rich R wrote:
waiting for TAC to explain a WLC crash on 17.9.5 seen in lab.

---

Same boat. 

We have a 9800-L that crash and went into a state we called "dark mode":  The console port was unresponsive, the distribution ports were dark.  Because it was not responding to anything, we had to cold-reboot the WLC.

[Rich R]

Ours (9800-80) just did SSO switchover so there is a system report file for them to decode - but suspiciously taking longer than it should ...

[Rich R]

- Have you done a line by line comparison of the configs on the 2 WLCs?

- Are you running TAC recommended IOS/SMU/APSP as per the link below?

[akcampbell]

I've not gone through each line between the two, just compared settings between the two in the GUIs.

As far as I know we're all good on the IOS side and respective patches and service packs.