Wireless Access

Hi,

 

I am looking at fine tuning the RTS/CTS settings at a site as there seems to be performance issues related to the wireless.

 

I see this parameter is in the ssid-profile, but was wondering if this should be change on all ssids at the site, or just the one in question?

 

Thanks

I would tell you that is probably the last thing you would ever change to solve a performance problem.  I personally would not touch it.

 

I would try every last suggestion in the thread here:  http://community.arubanetworks.com/t5/Community-Knowledge-Base/802-11-RF-Troubleshooting-Best-Practices/ta-p/20410 before editing the rts/cts settings.

 

 

Appreciate that, but the issue is bad enough to start considering all options.

 

A very high level of retries, 10-30%.  The clients are showing 1-15 % ping loss.  Interference index on APs is fine though.

 

It is happening with particular clients on a-HT. I've tweak a few parameters here and there and will have to wait on the testing tomorrow.  Latest driver for the clients actually makes things worse.

 

Next step is packet capture from an AM, and maybe from the controller at the same time, but any other suggestions appreciated.

 

Client:  HP thin client t5740e (Atheros AR5009  driver ver: 9.2.0.225)

 

 

 

 

Is Airwave in the environment?

How many Clients on one AP?

Are there other interferers?

Did you do a "show ap arm rf-summary ap-name" to see all the parameters on that AP?

          show ap arm rf-summary ap-name "AP125"

Channel Summary
---------------
channel  retry  phy-err  mac-err  noise  cov-idx  intf_idx
-------  -----  -------  -------  -----  -------  --------
161      0      0        0        0      0/0      0/0//0/0
1        0      0        0        91     5/0      29/6//0/0
48       0      0        0        0      0/0      0/0//0/0
165      0      0        0        0      0/0      0/0//0/0
5        0      0        0        89     0/0      11/19//0/0
6        0      0        0        85     0/0      11/19//0/0
7        0      0        0        84     0/0      0/24//0/0
11       0      0        20       93     7/0      26/2//0/0
149      0      0        0        0      0/0      0/0//0/0
36       0      0        0        0      0/0      0/0//0/0
153      0      0        0        0      0/0      0/0//0/0
40       3      0        2        89     11/0     21/0//0/0
157      0      0        0        0      0/0      0/0//0/0
44       0      0        0        0      0/0      0/0//0/0
HT Channel Summary
------------------
channel_pair  Pairwise_intf_index
------------  -------------------
1-5           65
7-11          52
149-153       0
36-40         21
157-161       0
44-48         0

Interface Name           :wifi0
Current ARM Assignment   :40/21
Covered channels a/g     :0/0
Free channels a/g        :9/0
ARM Edge State           :enable
Last check channel/pwr   :2d:23h:7m:38s/2m:22s
Last change channel/pwr  :2d:23h:7m:38s/2d:23h:7m:38s
Next Check channel/pwr   :0s/8m:4s

Interface Name           :wifi1
Current ARM Assignment   :11/21
Covered channels a/g     :0/2
Free channels a/g        :0/1
ARM Edge State           :enable
Last check channel/pwr   :1d:8h:50m:59s/6m:9s
Last change channel/pwr  :2d:18h:31m:7s/1d:15h:50m:11s
Next Check channel/pwr   :0s/24s

 

On a single client, what is the output of "show ap association client-mac" to see the client retries vs. the channel retries and error rates, as well as the noise?

         show ap association client-mac 00:90:4c:02:00:da

Flags: W: WMM client, A: Active, K: 802.11K client, B: Band Steerable

PHY Details: HT: High throughput; 20: 20MHz; 40: 40MHz
             ss:  spatial streams

Association Table
-----------------
Name   bssid              mac                auth  assoc  aid  l-int  essid    vlan-id  tunnel-id  phy          assoc. time  num assoc  Flags
----   -----              ---                ----  -----  ---  -----  -----    -------  ---------  ---          -----------  ---------  -----
AP125  00:1a:1e:20:82:e0  00:90:4c:02:00:da  y     y      2    10     CatchMe  1        0x108a     g-HT-20-1ss  32m:29s      1          WAB

00:90:4c:02:00:da-00:1a:1e:20:82:e0 Stats
------------------------------------------
Parameter                            Value
---------                            -----
Channel                              11
Channel Frame Retry Rate(%)          0
Channel Frame Low Speed Rate(%)      0
Channel Frame Non Unicast Rate(%)    0
Channel Frame Fragmentation Rate(%)  0
Channel Frame Error Rate(%)          2
Channel Bandwidth Rate(kbps)         0
Channel Noise                        92
Client Frame Retry Rate(%)           0
Client Frame Low Speed Rate(%)       0
Client Frame Non Unicast Rate(%)     0
Client Frame Fragmentation Rate(%)   0
Client Frame Receive Error Rate(%)   0
Client Bandwidth Rate(kbps)          0
Client Tx Packets                    1965
Client Rx Packets                    764
Client Tx Bytes                      210807
Client Rx Bytes                      462463
Client SNR                           34

 

 

 

yep, doing all that.  Many clients showing things not so good.

 

About to do some captures from the AP and session mirroring on the clients.

 

So back to my original question, if all options exhausted, then what would you suggest for RTS/CTS?

 

The options in the clients driver are basically nil.

 

:smileyhappy:

I wouldn't change that.  I would do a packet capture first, to find out what is really going on.

 

 

How are we sure it is a wireless/RF issue and not a networking or processing issue at the controller. We have seen incidents where high packet loss was experienced but it was the datapath at fault. I'm only mentioning this in your quest to explore all options.

RF issues can definitively spotted with a packet capture.

 

If a regular packet capture does not reveal anything, a controller performance issue would accomplished with TAC in conjunction with a packet capture.

 

What is common to both scenarios?  Yes, a packet capture.

 

Thanks guys.  I don't think it is RF, but I know this site to be a challenging rf environment, so was looking at that side first.

 

Have done the AP capture, but it didn't work at first.  The stream was in the datapath with 'D' and denied.  The TAC guy couldn't see why and we tried a few things with no luck.  Eventually got the customer to plug directly into controller and on same subnet as APs, which then worked.  The support guys will hopefully find out why that was the case.

 

Was doing some session mirroring on some clients to another laptop, but I couldn't see those streams in the the datapath either, but could see it hitting the acl.  Am I correct  in assuming that I should see something in the datapath for that session-mirror-destination ip ?

 

I'll get a capture from the controller port.  Unfortunately, that is on a port channel so have to pull one leg out to get all the traffic.  Would be great to be able to capture from a virtual interface like vlan or port channel instead of just a physical.  Sounds like a feature request to me though. :smileywink:

If you/they are seeing D flags in session table but don't know why, I would update all the acls/policies so that the deny statements (as well as the implicit deny statement) are set to log. This way "show log security" would uncover which policy these denies are hitting.

It was for an AP packet capture, so no acl as such.  I've just seen it again, so it appears to be a bug.

High number of retries.

 

Did you do a packet capture in the Air?  Who is retrying and why didn't the first packet go through in the first place?  Could you have a hidden node issue?  What is the utilization on the channel at that time?  How many devices can be seen on that channel when it happen?  What is the noise level measured by the AP on that channel?

 

 

Yes, packet capture in the air looks fine, though slightly higher percentage of retries compared to other sites, but everything else is the same.

 

All the RF stuff looks fine, and there are no issues on other device types.  I'm beginning to exhaust all options and starting to think that there is a manufacturing defect in this run of devices.

 

More testing to follow.

 

:smileyfrustrated:

Well it turns out these devices were advertising 2 stream capability when they only had 1 antenna.  Configured the ht-ssid-profile to have supported mcs index of 0-7 only and performance (raw download test) improved 10 fold.

 

Particularly difficult to troubleshoot as these devices are at other sites with no issues......same hardware, driver, software and same Aruba configuration.

 

For reference, they were HP thin client t5740e with wireless Atheros AR5009.  Driver version 9.2.0.480 seems to turn off the device advertising 2 stream capability, but changing the ht-ssid-profile to mcs 0-7 is an effective workaround.

 

Will make a seperate post regarding this with different title.

Just for information, but to disable RTS / CTS from network just put a value above 2346 in the RTS Threshold?

Thx

---

@Rafap wrote:

Just for information, but to disable RTS / CTS from network just put a value above 2346 in the RTS Threshold?

Thx

---

DO NOT touch it, at all.