QoS on MS350 Switches

[Atags]

Hi,

Is there anyway or logs debugs etc... to see if Meraki is doing QoS?

Basically I have MS350 switch stacks on each floor and Cisco IP 8851 phones connected to them.

Phone VLANs are 118,119,120,121 per floor respectively.

The phones are on its on separate dedicated POE link and vlan meaning PC's do not connect in the back of them. 

The MS 350 switches are setup in L2 fashion and are not doing any L3. L3 and SVI's are performed on our Cisco Catalyst Core switch.

Each stack connects back to the Core Catalyst switch over port-channels using 2 links (20G total).

I noticed that maybe once a day or couple times a week some calls are becoming 1 way audio.

On the Core, i have auto default QoS service policy's that are on the trunk links between floors. 

 service-policy input AutoQos-4.0-Trust-Cos-Input-Policy

 service-policy output AutoQos-4.0-Output-Policy

But when i look at the command to see if anything is matching, nothing seems to be matching, it all seems to be in the class-default queue. 

So back to my question, are their any way I can see the QoS is being trusted correctly like in CLI Cisco world to view if stuff is actually working?

Is at least the Meraki side look ok?

I have attached some photos of the settings i have on the Meraki's. 

These are the 4 voice vlans. 

This is a sample phone port. I did not put the vlan under the voice vlan since the PC is not piggy backing off the phones as mentioned above. 

Thank you.

[ww]

no u can 't.

but one-way audio is most time not a qos problem.  

did you take packet capture to make sure the meraki receives dscp ef values from the phone and sending them on the uplink to you catalyst?

[Atags]

This was from the IP phone port itself connected to a active phone call on the Meraki dashboiard

...

Network policy Subtype (2)
Application type [voice] (0x01), Flags [none]
Vlan id 4095, L2 priority 5, DSCP value 46

...

...

Network policy Subtype (2)
Application type [voice signaling] (0x02), Flags [none]
Vlan id 4095, L2 priority 4, DSCP value 32

...

I tried doing a 10 second port capture on the trunk links but it was so much data that dscp didnt show up in the search. I enable whole ball of wax as I am not sure which would show the dscp values. Low wasn't showing it. 

Does this seem accurate so far and correct DSCP settings coming into the port for cisco IP Voip?

[ww]

dscp looks fine.

your phone is hardcoded in vlan 4095?  or does it report that vlan because you did not set a voice vlan. 

[Atags]

No it's vlans 118-121depending on floor.

I saw that as well and was wondering why that is showing up? seems like it's a VLAN number after the standard ones firfCisco 1 thru 409 for if that means something?

Thanks

[ww]

i would try set voice vlan the same as data on a voiceport and bounce that port connected to (current unused ) phone. and capture Again.

[Atags]

You mean put voice VLAN 118 for example in the voice VLAN section in the port edit settings, instead of the data VLAN section and see what happens?

Thanks

[Atags]

Hi, Ok please see my tests below;

Remember they are not piggy backing any PCs off phones.

The Phone ports are set up as access and not trunk

Before port settings

Before a packet capture from a live phone call

...

 Network policy Subtype (2)
     Application type [voice] (0x01), Flags [none]
     Vlan id 4095, L2 priority 5, DSCP value 46

...

After port settings (999 is just a dummy vlan)

After change now a packet capture from a live phone call

Network policy Subtype (2)
     Application type [voice] (0x01), Flags [Tagged]
     Vlan id 118, L2 priority 0, DSCP value 46

Any thoughts on this now and why it changed it looked like?

Also why did the L2 priority change also?

Does this still look right for QoS?

Thanks