Cisco 3560 possibly stopping traffic to VLAN?

[Volteurope]

We have a remote office that has a 3560 configured with VLAN 1 for DATA and VLAN 2 for VOICE, a router connected to a leased line and a media gateway connected to ISDN.

Our IP phones are set to VLAN 2 and route internal calls through the router and external calls through the media gateway, randomly they will drop the incoming data for both internal and external calls (ie: Users in the remote office cannot hear anything but external or HQ internal users can). Also control signals to the phones dont make it through from HQ.

I am not sure if this is an Avaya phone issue or Cisco switch issue but as all the phones suffer at the same time it would appear switch related, it seems like the switch holds up or drops traffic from VLAN1 to VLAN2 but only occastionaly.

Does anyone have an idea what might cause this?

Thanks,

Simon.

 

[DallasBPF]

sounds like a QoS issue, do you have QoS set up on your switches and routers? also, during this outage time when this occurs what does the network look like usage wise?

because if you do not have QoS set up and the there are mass amounts of data being sent through the network, your voice will start to die...

------------------------------------
Dallas, Texas
Telecommunications Tech
CCVP, CCNA, Net+

CCNP in the works

 

[Volteurope]

QoS is setup on our routers but not the switch, I am fairly new to QoS so am not sure how to set it up on this switch.

The network traffic is normally quite minimal and I cannot see any bandwidth spikes aroud the time the phones have issues. We currently only have 4 users in this office and they are not heavy traffic users.

Any tips for QoS would be good, then I could rule this out.

Thanks,

Simon.

 

[vipergg]

I would go to the cisco site and lookup up how to setup auto qos on the 3560.

 

[NetworksMD]

Since this is only a 4 person office, I am assuming there is not a lot of inter office calling. If there has been are those calls being interupted as well?

How is the voice traffic from the media gateway being routed externally or to HQ? Does the MG have any local trunks? If the route is back out through the router and onto the leased line, and the leased line is carrying all traffic, you may be experiencing congestion on this line. It might even be congestion at the far end router since it seems like traffic coming to the remote site and not going to it. Depending on the QOS and WFQ settings these packets may be getting dropped or queued (which isn't good for real time data like voice).

 

[Volteurope]

There is no inter office calling as there are only 4 people in the office so I am not sure if its affected, the media gateway is connected through the router for internal signalling to HQ and through local ISDN lines for external call processing.

The phones use the Media Gateway for external calls and talk directly to each other (via the router) during internal calls, both the internal and external calls are affected and both use different communiction routes.

We have QoS on the Routers and MPLS network and I am looking at QoS settings for our switch, however we have a larger setup in our HQ which also does not have QoS and we have no issues with IP communications.

 

[vipergg]

If there are only 4 users then QOS is probably not an issue unless someone is occassionaly starting someone is starting a heavy download or something of that nature. Check all connections on the switch for any errors , you should see none if everything is set correctly. I have not seen any problems whatsoever in the 3560's we use . Also look at your router bandwidth on the line . If you only have say a T1 and if one or 2 people are doing a download then even QOS isn't going to help. Bandwidth is the ultimate QOS .

 

[tanderson3733]

If you only have say a T1 and if one or 2 people are doing a download then even QOS isn't going to help. Bandwidth is the ultimate QOS.

I don't agree. That's exactly what QoS is designed for, prioritizing one class of traffic over another, in this case voice traffic over downloads. If his routers are set using AutoQoS default EF class (Avaya's DSCP = 46) and even only 20% BW guarantee then they should be good.

Throw bandwidth at it, and you've added cost (recurring) and the users will just download more, ending up with the same problem.


Anyway, the original issue sounds less like a bad configuration and more like a physical hardware problem somewhere...switch, router, cabling etc. There could also be a dirty circuit...you can look at errors on the T1 interfaces for a starting point.
Also take a look at the Media Gateway...what model is it (G350, G700 etc)? I would be curious to see if it unregisters from the MGC when you have this problem.

 

[burtsbees]

Well, to eliminate the phones being the issue, you can do an extended ping -n for 12 hours and look for drops. That would point to a hardware issue in the 3560. You can also create different access ports in the switch and connect the vlans to different switchports, and change cables while you're at it. Do incoming AND outgoing drop at the same time? What BW do you have at the gateway and router? Leased line like frame could be as little as 56K, and ISDN at 64K. Perhaps this is what viper means by BW being the ultimate QoS...
Post a sh run from the 3560 also, and sh int from switchports in vlan2. Let's look for interface resets, late collisions, giants, runts, etc.
You may also want to run a sh tech through Cisco Output Interpreter, if you have a login...

Burt

 

[Volteurope]

Many Thanks for all your suggestions, I am working my way though various test that have been suggested.

The MG in the remote site is a G350 and does not lose contact with the 8300/G650 in HQ. QoS is in place on the Routers and 10Mb leased lines at both the Remote site and HQ, just not on the switches.

Yesterday I changed a phone to VLAN 1 and updated the DHCP settings and Network Map to allow it to function. This morning the issue occured again but did not affect the phone i changed to VLAN 1.

 

[burtsbees]

Post a sh run from the 3560...

Burt

 

[Volteurope]

Building configuration...

Current configuration : 4787 bytes
!
! Last configuration change at 15:36:35 GMT Wed Jan 21 2009
!
version 12.2
no service pad
service timestamps debug uptime
service timestamps log uptime
no service password-encryption
!
hostname *************
!
enable secret *************
!
no aaa new-model
clock timezone GMT 0
system mtu routing 1500
vtp domain *************
vtp mode transparent
ip subnet-zero
ip routing
ip domain-name *************
ip dhcp excluded-address 10.2.6.200 10.2.6.254
ip dhcp excluded-address 10.3.6.200 10.3.6.254
ip dhcp excluded-address 10.3.6.1 10.3.6.99
!
ip dhcp pool LondVoicePool
network 10.2.6.0 255.255.255.0
default-router 10.2.6.254
option 176 ascii "MCIPADD=10.1.5.12,MCPORT=1719,HTTPSRVR=10.1.1.80
"
option 242 ascii "MCIPADD=10.1.5.12,MCPORT=1719,HTTPSRVR=10.1.1.80
"
lease 14
!
ip dhcp pool LondPool
network 10.3.6.0 255.255.255.0
domain-name *************
default-router 10.3.6.254
dns-server 10.3.1.10 10.1.1.7
option 176 ascii "MCIPADD=10.1.5.12,MCPORT=1719,HTTPSRVR=10.1.1.80
"
option 242 ascii "MCIPADD=10.1.5.12,MCPORT=1719,HTTPSRVR=10.1.1.80
"
!
ip dhcp pool LondStatic
host 10.3.6.99 255.255.255.0
client-identifier *************
client-name *************
default-router 10.3.6.254
dns-server 10.3.1.10 10.1.1.7
domain-name *************
!
ip dhcp-server 10.2.6.254
!
!
!
no file verify auto
spanning-tree mode pvst
spanning-tree extend system-id
!
vlan internal allocation policy ascending
!
vlan 2
name LondonVoiceVLAN
!
interface FastEthernet0/1
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/2
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/3
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/4
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/5
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/6
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/7
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/8
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/9
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/10
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/11
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/12
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/13
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/14
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/15
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/16
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/17
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/18
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/19
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/20
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/21
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/22
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/23
description Backup Router
switchport voice vlan 2
spanning-tree portfast
!
interface FastEthernet0/24
description Primary Router
switchport voice vlan 2
spanning-tree portfast
!
interface GigabitEthernet0/1
!
interface GigabitEthernet0/2
!
interface Vlan1
description London Data VLAN
ip address 10.3.6.254 255.255.255.0
!
interface Vlan2
description London Voice VLAN
ip address 10.2.6.254 255.255.255.0
!
ip classless
ip route 0.0.0.0 0.0.0.0 10.3.6.253
ip http server
!
control-plane
!
!
line con 0
line vty 0 4
password *************
login
length 0
line vty 5 15
password *************
login
length 0
!
end

 

[VinceWhirlwind]

What about your HQ 3560 config?

Also, check your router QoS config is working by checking that the traffic on the 10Mb link is tagged correctly and that the tagging is being honoured correctly.



Can you clarify what the remote site media GW actually does during a normal call?
ie, do the remote phones register to their local GW, or to the HQ GW?
Do the phones send traffic directly phone-phone or does the remote GW come into it somehow?

The fact that your "control signals" are going from HQ to remote phones tells us a bit, and if it is a fact that they also are failing indicates it's absolutely not a QoS issue but a service issue. (Time-outs on TCP traffic is usually 3-30s, and 3s is 10x longer than normal VoIP traffic can handle anyway).

As Burtsbees has recommended, monitor the connection between sites for connectivity lapses.
Also monitor connectivity to & between each GW.