Wireless Access

Hey guys, 

I was wondering if anyone encountered a problem wherein, I already uploaded a floorplan and placed the APs on VisualRF however heatmap or doing a survey does not work. All the clients show up on 1 spot in the map. Even viewing the AP's rf environment does not show on the floorplan.

Thanks

See if there is anything in the logs in System> VisualRF Engine.

When you left click on an AP or client, are there any errors on their detail pop up?  Is the status reporting 'down'?

If you right click on an AP, is the option showing as 'simulate failure' or 'disable failure simulation'?

If you have command line access, check the VisualRF log for crashes.  Log located at /var/log/visualrf/visualrf.log

We typically see issues with heatmaps if the catalog doesn't contain the devices you're working with.

What version of AMP are you running?  What type of devices / firmware do you have place on your floor plans?

Current version of AirWave available is 7.5.6.

If you're on the latest, it may help to open a support case so that they can take a look at the backend to see if there are any errors.

When you left click on an AP or client, are there any errors on their detail pop up?  Is the status reporting 'down'?

I get two errors for AP it is "AP error: radio11ng is invalid" for the Clients it is "Signal data from device is invalid. the client has not been located". and status is saying UP.

If you right click on an AP, is the option showing as 'simulate failure' or 'disable failure simulation'?

 No, I am not using visualrf plan and not simulating anything.

If you have command line access, check the VisualRF log for crashes.  Log located at /var/log/visualrf/visualrf.log

I was able to access the logs, there are some logs that are significant. Example logs are below.

Logs:

2012-10-30 04:33:07,720 ERROR AirBusMs[1]  com.airwave.model.impl.APConfigProcessorImpl Catalog definition not found for manufacturer[Dell] model[PowerConnect W-AP93]
2012-10-30 04:33:07,720 ERROR AirBusMs[1]  com.airwave.model.impl.APConfigProcessorImpl Catalog definition not found for manufacturer[Dell] model[PowerConnect W-AP93]

2012-10-30 04:35:37,462 WARN  Grid Builder com.airwave.rf.SignalServiceImpl Survey discovering radio not found for [18.1]  at signal[-43]
2012-10-30 04:35:37,462 WARN  Grid Builder com.airwave.rf.SignalServiceImpl Survey discovering radio not found for [12.1]  at signal[-38]

We typically see issues with heatmaps if the catalog doesn't contain the devices you're working with.

What version of AMP are you running?  What type of devices / firmware do you have place on your floor plans?

Current version of AirWave available is 7.5.6.

We are using version AMP 7.4.8

I have a question too, what does the catalog do? how do I add devices into the catalog?

This is what I was suspecting.  The catalog repository maintains all of the values used by VisualRF to calculate location based on device / radio type.  I'm going to spin up an AMP on 7.4 to see what the catalog repository had, but I recall that the AP93 was released post 7.4, so you're quickest solution might be to upgrade to 7.5.6.

Correcting my last reply.  The entry is in 7.5, but not 7.4.  Once you upgrade to 7.5.6, you should be okay.

Awesome we are in the process of getting the image file for 7.5.6. Should we get it from Aruba or from Dell since we have a Dell branded AMP instance.

Also, may I ask where or what documentation did you find that AP93 is not in 7.4 and is in 7.5?

Thank you so much for the help

If you're running a Dell branded AMP, you should always go to Dell for the upgrade image.  If you try to use the Aruba image, it will error out.

I didn't look in any documentation, I installed 7.4.8 and 7.5.6 on two AMPs, and then compared their catalog repository files.  Though typically the release notes will say when support for a device was added.  It's worth noting that since Dell renames the type string, support for Dell OEMs sometimes occurs a patch or two after the Aruba patch is made.

Awesome thank you!.

I have another issue that I wonder if you can help me with, so my aps that are in Airwave have the error below.

This Device is in monitor-only mode. 

Configuration read from device at 10/30/2012 10:09 AM

Configuration: Error (Too many errors fetching existing configuration)

And from the event logs we see the error below:

Tue Oct 30 10:20:01 2012 System Device Dell PowerConnect XXXX P3SW05 Configuration verification: failed to read configuration from device
Tue Oct 30 10:20:01 2012 System Device Dell PowerConnect XXXX P3SW05 Status changed to 'Error fetching existing configuration'
Tue Oct 30 10:20:01 2012 Dell PowerConnect W-AP93 XXXX Configuration status changed to 'Too many errors fetching existing configuration'
Tue Oct 30 10:20:01 2012 System Device Dell PowerConnect W-AP93 XXXX Configuration status changed to 'Telnet/SSH Error: (pattern match timed-out) in password prompt: '

I want it to audit the configs of each APs, I was wondering if this is related to it not being in the catalog either?

1. Sanity check group settings

Go to Groups -> select group controller is in -> go to Basic tab

Make sure that the SNMP version is set correctly

2. Sanity check the credentials AMP has for your controllers:

Go to Groups -> select group controller is in -> select controller -> go to Manage tab -> View Device Credentials

Confirm that the credentials given work from the AMP CLI

If you're on SNMPv2c, then command will look like:

# s2w ip.of.controller community_string oid flags

ex: # s2w 10.10.10.10 public sysName

If you're on SNMPv3, then command will look like:

# s3w id_of_controller oid flags

     where id_of_controller is AMP's id that you can get from the URL

ex: if URL = amp.ip.address/ap_monitoring?id=1234

# s3w 1234 sysName

Also check telnet/ssh from AMP to the controller from the AMP's CLI:

# ssh admin@controller

enable and do a # show running-config

If all of the above is possible with the credentials you have on the manage page, then it should all work.  Try going to audit page and running the audit.  If there's still issues, make sure there's not more to the errors.

# /var/log/ap/(url_id)/telnet_cmds

where url_id is gotten the same way we got the controller's id above.

So I checked the controller and verified the V2c snmp community string.

I also checked the AMP and verified in the the device credentials of the control. It is using the same SNMP v2c string. 

Question: But doesn't that just tell me what snmp string is of the controller? Where in AMP do you configure to tell AMP what community string to use if it wants to poll the AP?

Also I verified that AMP cannot ssh to the controller using the CLI, I think it is a network issue because I can't ssh to it either.

Question 2: Does the AMP need to SSH to the controller?

Thanks

1. If you look at the APs on a controller in AMP, you'll notice that AMP doesn't show the 'view device credentials' link for the APs on their manage page.  It's because Aruba APs are thin APs that do not retain anything but the config that the controller gives them.  Therefore, the controller is the main source for all controller and AP related data (and by extension client data).

2. AMP requires SSH or Telnet to controller to get the running-config for Audit comparison.  Without this, AMP is not going to be able to audit the devices properly.

Thank You.

So updating to 7.5.6 fixed the heatmap problems with the AP93s. As far as the configuration error problems we are getting, I still am trying to figure out why we can't ssh to the controller.

Problem right now is I cannot see any clients show up in visualrf. APs statuses are all OK.

These are the errors I get.

2012-11-01 12:11:36,529 WARN  HTTP[2]      com.airwave.svg.APIFacadeImpl Parse heatmap xml request: 0 aps, 6 radios.
2012-11-01 12:12:12,067 WARN  HTTP[2]      com.airwave.svg.APIFacadeImpl Parse heatmap xml request: 0 aps, 6 radios.
2012-11-01 12:16:23,297 WARN  HTTP[1]      com.airwave.svg.APIFacadeImpl Parse heatmap xml request: 0 aps, 12 radios.

These aren't actual errors, only parser warnings.  Try to poll the controller again.  Make sure that you can drill down from the controller, to the AP of association, to the end-user client.  If all that is successful, try to view the client.xml file:

http://your-amp-name/visualrf/client.xml

Run a search for the client's username or MAC address.

If you see the client there, verify that the x and y coordinates have a value.

If there's an x and y coordinate, then take a step further to verify it on the floor plan by getting the site.xml.

To get site.xml:

Go to VisualRF

Navigate to the floor plan that the has the AP of association placed

Click on the map so that the flash frame is the active window

Press ctrl+x+m+l at the same time to launch the site.xml.

From there, you can search for the client again, to double check that the site.xml matches the client.xml.

If you've done this and are still having issues, open a support case and a support engineer can help gather more info to troubleshoot.

I can see all the APs and clients in the AMP.

However when I check the client.xml, there are only like 5 clients there from the 70 clients that I see in the AMP.

Yeah, that's along the lines of what I was imagining.  At this point, it's harder to troubleshoot this over text without a visual of the system.  Next step should be to open a support case.

One thing you might try is to restart the AMP services (it's a possibility that the thread that sends VisualRF data may be in a bad state).  To do this, go to System -> Status -> select 'Restart AMP' at the bottom of the page.  Best to do this during a maintenance window since it will stop AMP from recording data for about 10-15 minutes as it restarts processes.  Forcing the processes to restart may clear out or restart any services that may be stalled or in a bad state.

I tried restarting the AMP, and it briefly showed the clients being located in visualrf. However, after a couple of mins they all dissappeared.

It was worth a try.  The support engineer will may need to do it again to capture some data and possibly when the process halts/crashes.  The main branch of traffic is going to be over airbus, but will need a live connection to troubleshoot.  Please do open a support case.

This sounds like a problem which I also was having.  The support engineer gave me a replacement component file for visualRF and with it in place on AMP, VisualRF is working fine -- I see clients and AP details correctly now.