[cisco-voip] Phones change to DECEASED ....

[Mario Klar]

Base information:

**Cisco Call Manager Express 4.2
Integrated Services Router 2851 (revision 53.51)
IOS Version 12.4(11)XW2
VLAN 1: Daten  VLAN 2: VoIP
Phones in use: 7940,7941,7960,7970,7921

*Description of the problem:
*In CME ephones change there state to DECEASED after a while without without
a recognizable cause.

In this state the concerned phones repeat the following procedure again and
again:
- configuring IP
- configure CM-list
- opening <ip-cme> 6-10 sec
- configuring IP
- configure CM-list
- opening .......... and so on

In this state the phones get there IP information per DHCP correctly
(IP,SM,GW,TFTP,CME)
A restart of the phone doesn´t bring it back alive.
Unplug power for a while and plug again brings us to the procedure above.
A restart of all phones using telephony-services "restart all" takes all
phones to state unregistered and descripted procedure
Deleting a concerned ephone and create new doesn´t change anything
The only way to geht the phones back in registering procedure is to restart
the router. What causes the unregistration of all phones - bad in working
hours

After restart all phones register fast and correct. Strange to say but up to
one hour (waited at customer site for about one hour testing phase) after
restart phones register absolutly normal when unplug from power.

Next morning....some phones are in state DECEASED again. If I now take a
phone from power and plug it again...the procedure above starts.
Only way to geht the DECEASED phones back for work is....restart the
CME-Router.

What else tried:
Already said: restart, unplug power, restart all, delete and re-create
ephone

- changed the DHCP Option 150 and ip source address back from 172.20.221.199to
192.168.100.1
  (we had changed this out of the 192er VoIP-Subnet because of the
registration over WAN with IP Communicator for example...worked fine)
  --> no results, same behaviour

- ckecked for error messages using debug dhcp, debug ephone register, debug
tftp events ...... but all is normal. In descripted error procedure no
messages appear.

- checked the show ephone output. All phones with state deceased have "TCP
socket:[-1]", don´t really know what this could mean

- changed the keepalive in telephony-service from 30 to 3000
  --> the same behavior after weekend. First phone goes deceased after 11
keepalive, next after 26 keepalive counters, next after 55 and so on.


What else can I try? What else couls be the reason that the phones go
deceased?
Maybe hardware failure?
There was an power loss two days before failure occurs for the first time.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://puck.nether.net/pipermail/cisco-voip/attachments/20080401/f5167233/attachment.html