I have a 1638P and have the following set up:
Bsmt24Port#sho qos cos
CoS Priority : 0 1 2 3 4 5 6 7
Priority Queue : 1 1 2 2 3 3 4 4
Bsmt24Port#sho qos dscp
DSCP-CoS map:
DSCP: 46
----------------------------------
CoS: 6
Bsmt24Port#sho qos que
Queue-type: strict-priority
Expedite queue: n/a
Strict-priority levels:
qid - priority level
1 - lowest
2 -
3 -
4 - highest
Bsmt24Port#
Bsmt24Port#sho run int gi 0/15
Building configuration...
!
!
interface gigabit-switchport 0/15
spanning-tree edgeport
no shutdown
switchport access vlan 2
switchport voice vlan 3
qos trust cos
no lldp send-and-receive
I am seeing low MOS scores for individual calls throughout the day, with all of them showing delay for the reason of the low score. There is no pattern as to when this happens, and they are not a data intensive business. A wireshark capture confirms EF tagging
Phones are on VLAN 3, PC’s are VLAN 2 and plugged in the back of the phones.
Everything else appears ok. Are there any built in tools on the 1638, that will show switch cpu/memory/and queue load?
Is this the only switch in your environment? If so, are you using "qos trust cos" on the uplink to your voice server (or the Internet if your phone system is cloud hosted)? If not, are you using "qos trust cos" on both sides of every switch trunk?
Thanks for the reply Phil. I am using qos trust on all switchports and trunks. The issue turned out to be a problem with a specific version of software on the VoIP phone system