Re: 2802 AP interface usage

AK002 · ‎10-05-2017

Hi All,

We are having one query, One of our 2802 model AP which is integrated with 5520 controller in 8.2.151.0 version having some client disconnection issue. Now its random on some other AP's in same model.

AP's is configured in LAG and its showing packets in both gi0 and gig1 interfaces, But other Ap's we can see packets only in gig o.

How it works normally, Most of the AP showing only in gig0, we suspect the disconnection is due to the pac mismatch.

Regards,

Arun

Flavio Miranda · ‎10-05-2017

AP doesn't support LAG. WLC does.Is it the AP in Local Mode or Flexconnect?

This AP has two ethernet interface, which one did you link to the switch?

About client disconnection, you should analyse coverage first. If they have good signal level and still disconnects, then, you can go to a next level.

AK002 · ‎10-05-2017

We are having 2802I model AP's which support LAG and switch end its LACP.

Coverage is fine. Both the links(Port1 and Aux) were connecting to Siwtch end via etherchannnel in LACP.

Auto Link Aggregation (LAG) support

802.3ad (LACP) compliant, allowing both Gigabit Ethernet interfaces to automatically LAG, increasing overall throughput to the access point.

Flavio Miranda · ‎10-05-2017

that´s nice!

Could you share with me how did you setup LAG on it?

And, if you put two interfaces as an etherchannel, why are you find it weid to have traffic on both interfaces?

Leo Laohoo · ‎10-05-2017

Client disconnection has NOTHING to do with the APs in LAG or not. This information is totally irrelevant.
If the clients face constant disconnection, then focus needs to be on the clients as well as the WLC.
About the clients facing disconnection, find a pattern:
1. What kind of OS? Windows? Android? Chrome? iPhone? Mac?
2. If Windows, what is the exact make and model of the wireless NIC and and what is the exact wireless NIC driver version?
3. During the time of the observed disconnection(s), what was/were the clients doing? Moving around? Stationary? Sleep?

AK002 · ‎10-06-2017

Hi Leo,

Issue suddenly coming to some radndom AP's, last week it was another set of AP's (same 2802) after reboot it started working, Now another set of AP's.

Why i mentioned Gig1 is traffic is all otherAp's which are working fine we see traffic only in gig0 but this issue Ap's we see packet in both gig0 and gig1.

During issue client are in idle state only not roaming. after disconnection it was connecting to far/nearby AP.

Mostly all are mobile devices.

Below are the show log from AP during the issue.

Oct 3 15:38:48 NCI: I1: channel map channels: in=1 cloned=1
Oct 3 15:38:48 NCI: I1: Requesting MonBand [5240, 5240] bw=20MHz/0 ant=0xbc
Oct 3 15:38:48 NCI: I1: Monitoring (cf=5240, span=20), RadioUsage=3%
Oct 3 15:38:48 NCI: I1:     dwell=20000us, update=1000ms, resBW=78126
Oct 3 15:38:48 NCI: CLEANAIR: Slot 1 enabled
Oct 4 13:48:35 kernel: [*10/04/2017 13:48:35.8487] invalid staid 55 received
Oct 4 16:11:47 kernel: [*10/04/2017 16:11:47.6163] invalid staid 96 received
Oct 5 07:37:59 kernel: [*10/05/2017 07:37:59.8463] invalid staid 39 received
Oct 5 07:57:50 kernel: [*10/05/2017 07:57:50.9156] invalid staid 20 received
Oct 5 10:36:45 kernel: [*10/05/2017 10:36:45.4981] invalid staid 240 received
Oct 5 13:19:37 syslog: Password for 'admin' changed
Oct 5 13:19:37 kernel: [*10/05/2017 13:19:37.6587] Password for 'admin' changed
Oct 5 13:19:37 kernel: [*10/05/2017 13:19:37.8338] In write handler 'run' for 'wired_port_enable :: Script':
Oct 5 13:19:37 kernel: [*10/05/2017 13:19:37.8338]   While calling 'port_manager.sec_eth_port_enable 1 0 0 false false':
Oct 5 13:19:37 kernel: [*10/05/2017 13:19:37.8338]     Error - wired1 is a LAG port
Oct 5 14:03:49 kernel: [*10/05/2017 14:03:49.0505] In write handler 'run' for 'wired_port_enable :: Script':
Oct 5 14:03:49 kernel: [*10/05/2017 14:03:49.0505]   While calling 'port_manager.sec_eth_port_enable 1 0 0 false false':
Oct 5 14:03:49 kernel: [*10/05/2017 14:03:49.0505]     Error - wired1 is a LAG port
Oct 5 14:37:12 NCI: CLEANAIR: Slot 1 down
Oct 5 14:37:12 NCI: I1: shutdownNci
Oct 5 14:37:12 kernel: [*10/05/2017 14:37:12.3732] DOT11_DRV[1]: set_channel Channel set to 52
Oct 5 14:37:12 kernel: [*10/05/2017 14:37:12.5170] wcp/dfs :: RadarDetection: timer[1] stopped.
Oct 5 14:37:13 kernel: [*10/05/2017 14:37:13.0993] change to DFS channnel-1, CAC for 60 seconds.
Oct 5 14:37:25 NCI: I1: openSensor(slot=1)
Oct 5 14:37:27 NCI: I1: SensorApp=1.15.4
Oct 5 14:37:27 NCI: I1: SensorHdw=1.2.3.0
Oct 5 14:37:27 NCI: I1: Hardware Radio Band = [4890, 5935] MHz, BW=150625
Oct 5 14:37:27 NCI: slot=1 mode=0 chanCnt=1 cw=1
Oct 5 14:37:27 NCI: chans: 52 0 0 0 0 0 0 0 0 0 0
Oct 5 14:37:27 NCI: I1: channel map channels: in=1 cloned=1
Oct 5 14:37:27 NCI: I1: Requesting MonBand [5260, 5260] bw=20MHz/0 ant=0xbc
Oct 5 14:37:27 NCI: I1: Monitoring (cf=5260, span=20), RadioUsage=3%
Oct 5 14:37:27 NCI: I1:     dwell=20000us, update=1000ms, resBW=78126
Oct 5 14:37:27 NCI: CLEANAIR: Slot 1 enabled
Oct 5 14:38:12 kernel: [*10/05/2017 14:38:12.8970] event-> CAC_EXPIRY_EVT
Oct 5 15:06:40 kernel: [*10/05/2017 15:06:40.8115] In write handler 'run' for 'wired_port_enable :: Script':
Oct 5 15:06:40 kernel: [*10/05/2017 15:06:40.8116]   While calling 'port_manager.sec_eth_port_enable 1 0 0 false false':
Oct 5 15:06:40 kernel: [*10/05/2017 15:06:40.8116]     Error - wired1 is a LAG port
Oct 5 15:27:46 kernel: [*10/05/2017 15:27:46.0917] invalid staid 30 received

Leo Laohoo · ‎10-06-2017

Since you've got 2800, try upgrading to 8.2.161.0 and see if the issue reoccurs.

AK002 · ‎10-06-2017

Hi Leo,

Can you go with 8.3.122.o version instead of 8.2 series.

Leo Laohoo · ‎10-06-2017

If you want to use 8.3 then use 8.3.130.0 instead.

AK002 · ‎10-09-2017

Ok i will upgrade and let you know the result.

palcantara · ‎12-18-2017

hi I look same issue on 2802e and same log error, post upgrade the issues fixed?

@AK002 wrote:

Ok i will upgrade and let you know the result.

junaidnaseer · ‎05-20-2018

Hi Arun,

Hows your subjected AP2802 disassociation with WLC solved ?