Nortel Phones drop out randomly

[Flintstone]

Hi subbusvce,

Even though your IP Phones are connected to external power the IP Phones I believe will also draw POE from the switch?

CheerZ

[Michael McNamara]

If the IP phones are rebooting then it is most likely a watchdog timer issue. If that's the case then it is most likely a QoS issue within your network.

You can read here for some background on watchdog timer issues (the article is around Nortel wireless handsets but the same principles apply - although the IP deskphones won't display anything about watchdog timer).

If the IP deskphone (and wireless handsets as well) fail to hear back from the voice infrastructure in a set amount of time the "watchdog timer" will expire and the IP deskphone will assume something bad has happened and will attempt to reboot itself to try and recover.

If you ask your telecom folks to review the logs on the CS1000 or the Signaling Servers they should find that the TNs are timing out and disconnecting. If they enable QoS alarms on the CS1000 they should start seeing some significant QoS alarms on the console assuming they have the "Zones" setup properly.

Was there any MOS testing ever performed on the network before you deployed the IP deskphones? What was the score?

Good Luck!

[subbusvce]

Flinstone, we have turned off the power on the ports (power never inline) on the switch..so its just external power alone

Michael, we dont have any qos set up at the moment. some one did say about this timer but the company that has installed the phones say there are no timer set at all..by default is there a timer set on these phones if so what is the time set ?

we have applied similar rollout on another site (but jsut a few phones..20 phones) where they work perfectly fine..no QoS here again.

if i have to apply QoS do you have any config that can be used on out setup

Thanks for your help

[beentherebefore]

What do you see in the switch logs?
If you are losing both computer and phone it would make more sense to look at the switch to me.
If it was just the phone rebooting and the computer not disconnecting I would suspect the connection back to your VOIP system may be involved. Nortel phones will reboot if they cannot see the pbx they are registered to, or fail over to the branch in local mode. 

[subbusvce]

Hello,

When the phone looses connection the pc connected to back of the phone looses connection also. i am not sure how to solve this problem because its been happenning for some time now...the strange thing is that we have deployed similar phones in another site (but just around 15 phones) and they seem to be fine...no rebooting...but the first site has around 200 phones which keep rebooting...the phones are the same in both place...


The phones go and register with a third party company and they have a seperate link just for the phone traffic...the LAN traffic would be same but the there is a seperate link for the WAN traffic..now any idea how i need to apply QoS for this...may be its the QoS thing thats affecting this problem...Micheal any thoughts on this ?

Thanks

[RMcEnery]

The Watchdog timer is 200 seconds. The RUDP cycle starts every 100 seconds. The RUDP cycle also has a 10 second watchdog in case it is unable to complete its process. The IP phone may restart anywhere from 110-200 seconds after a succesful RUDP cycle.

On the phones with Static IP, are their IP and Voice VLAN fixed on VLAN 20? I see VLAN 20 DHCP offers the "VLAN-A" string. This is only necessary in the default VLAN on the switch port. As soon as the IP phone receives it, it will restart the DHCP request on the listed VLAN(s).

I agree with Michael that lack of QoS could be at the heart of this. The office was constructed with separate 2950 switches for voice and data. Any time a PC was connected through an IP phone, and transmitted a print job or a file, somebody on the edge switch complained about interference on a call. We had no choice but to implement QoS and eliminate this scenario.

My customer is using Cisco switches and routers from the core to the edge. Avaya (Nortel) aligns the DSCP values for Control, 40 and Media, 46 with VLAN priority 6. Cisco calls this Priority, Class of Service, but they align the VoIP DSCP values with COS 5. It was easier to conform to the Cisco VLAN-DSCP mapping by adding it to the DHCP offer. I cleaned up the DHCP strings. The Data VLAN offers option 191 - "VLAN-A:x." and the Voice VLAN offers option 128 - "Nortel-i2004-A:<Server1 IP>:4100,1,3;<Server2 IP>:4100,1,5." and option 144 - "Nortel-i2004-B,eap=dis;lldp=n;cachedip=n;vq=y;vvsource=a;vcp=5;vmp=5;vlanf=n;pc=y;pcs=a;pcd=a;dq=y;dv=n;dp=8;st=y;pcuntag=n;igarp=n;srtp=n;bold=y;prov=<TFTP Server IP>;zone=<Zone ID>;". The customer has deployed a PXE BootP server for thin clients, I needed the "Nortel-i2004-B" string to make use of provisioning the 11xx phones.

While i2004-A string and i2004-B string are not recommended in the same DHCP offer, I kept the information paired down to what is necessary. I can succesfully use Full DHCP or Full Auto Settings on all sets: i200x phase 0/1 (B76) and 2 (DCN), 11xx (C7J) now updated to Unistim 5.2 (C8G) and 2033 conference phones on S77. No Touch Provisioning from the TFTP server is also working.

Implementing Quality of Service Policies with DSCP  [QoS Packet Marking] - Cisco Systems: http://www.cisco.com/en/US/tech/tk543/tk757/technologies_tech_note09186a00800949f2.shtml

Regards!

[Michael McNamara]

I honestly couldn't tell you where to begin.... you really shouldn't deploy VoIP unless you QoS fully configured across your entire network. You're only asking for a really large headache. I would go back and start deploying the proper QoS configurations on all your equipment. How big a pipe is this WAN connection to the Call Server? How busy is it? Why don't you setup a packet trace using WireShark... can you see issues in the trace? You don't have a loop somewhere in the network do you? Do you have Spanning Tree (and FastStart) enabled on all the edge ports at a minimum?

Good Luck!

[Gord Webster]

When the phone looses connection the pc connected to back of the phone looses connection also.

That would tend to indicate that the issue is somewhere with your data network.

[Michael McNamara]

If the IP phone looses connection to the Call Server/Signaling Server (or Aura Session Manager) the IP phone will reboot. Any reboot of the IP phone will cause the PC port to become temporarily disconnected as the built-in switch looses power and is reset. So regardless of what's causing the IP phone to reboot... even if you intentionally reboot the IP phone that will cause the device on the PC port to be temporarily disconnected from the network.

Good Luck!