Symptom: A router will run out of memory when SIP phones register
Conditions: Cisco 3911 phones are installed
Workaround: Disable MWI

Symptom: Customer running 12.4(20)T is seeing %ALIGN-3-SPURIOUS
Conditions:
Workaround: None
Further Problem Description:

Symptom: Unable to transfer calls directly to voice-mail using the "TrnsfVM" soft-key followed by a speed dial, "watch" or "monitor" button on CCME running 12.4(20)T IOS.
Conditions: Transferring calls directly to voice-mail using "TrnsfVM" soft-key followed by a speed dial, "watch" or "monitor" button
Workaround: Use: http://www.cisco.com/en/US/customer/products/sw/voicesw/ps5520/products_tech_note09186a00802ab979.shtml

Symptom: If final number of the sequential hunt group is the pilot number of another hunt-group, call does not go to the new huntgroup, instead the call is disconnected.
Conditions: Happened in simple SIP voice huntgroup configuraton
Workaround: N/A

Symptom: Periodical crashes on 2800 with CME features.
Conditions: When "callmonitor scan" is configured
Workaround: Turned off "callmonitor scan" n/a

Symptom: When trying to access the CCME gui only view page is displayed in the upper left hand corner. The router does not seem to continue to load the files in flash.
Conditions: We found that if you use " quotations " around the directory name configuration under telephony service the IOS stops loading the gui files as it puts its own " quotation marks " around the name so you wind up with double.
Workaround: remove the quotation marks.

Symptom: No voice path after connected if CME receives 183 with sdp and then 180
Conditions: reproducible
Workaround: called party only send 183 or 180

Symptom: When an ephone hunt-group is configured with 'present-call idle-phone', the ephone hunt-group skips over certain members of the hunt group.
Conditions: The problem is observed when members of the ephone hunt-group are monitored.
Workaround: Do not monitor the members of the hunt-group

Symptom: Skinny Msg Server process increases its memory usage eventually crashing the router
Conditions: SCCP devices trying to register while not in SRST mode

Symptom: With CME 7.0 Phones randomly reregister as another extension. When you restart the router, they register correctly again
Conditions: Running ios version 124-15.XZ1 or 124-20.T1
Workaround: downgrade to cme 4.2

Symptom: When using the TransVM (transfer to voicemail) soft key, the softkey does not operate as expected unless it is used before trying a Transfer.
Conditions: If a full consult transfer is attempted, and the called party elects not receive the transfer (but would like a voicemail), the caller who initiated the transfer cannot use the TransVM softkey because the functionality is missing. It does not permit the dialing of the voice mailbox extension (as it would had the full consult transfer not been attempted).
Workaround: The work around without upgrading to code which provides this fixed softkey feature is to have a dial peer which would transfer say to *### to get to the voicemail of extension ###.
Further Problem Description:

Symptom: The CME "IP Input" process will hold/leak increasing amounts of memory over time during an H.323 sccp video call.
Conditions: The leaking CME router is receiving packets from a version 2.0.2 CUVA.
Workaround: none.

Symptom: Customer has a remote branch that has it's own unity box (96 ports) and 30 IP Phones. When the WAN drops, all 96 voicemail ports and the 30 phones register to the CME as SRST gateway in a first come/first serve basis. The customer has manually created 11 phone(s), and put the vm-device-id command under them, to reserve the some ports for unity. The first time the WAN fails, the system works as expected. The first 11 ephones are used exclusively for voicemail ports. The phones and the rest of the voicemail ports race for the remainder of the regsiterations. The problem is that on the second WAN failure, the first 11 ephones will then allow an ip phone to register in that spot.
Conditions: WAN fails a second time, and there is no control over what device registers as what ephone
Workaround: CME allows for auto assign, but this breaks SRST's provision command.

Symptom: CCME port becomes stuck in a SEIZED state when integrated with Unity TSP and set to dialout MWI, which prevents any future MWI from being set through that port
Conditions: CCME integrated with Unity / Unity Connection running TSP 8.1(3) or earlier
Workaround: Reset the MWI port in CCME OR for Unity Connection: increase the timing interval to 1500ms by going to "Port-Group" and modify the "Delay between requests" setting, then restart Unity Connection OR for Unity: open the \CommServer\InitLib\Cisco0002.ini file in Notepad, then add the following lines at the end of the file: [Configuration] MinimumMWIRequestInterval=1500 Save your changes, then restart Unity Additional Information By increasing the timer to 1500ms, the TSP will wait 1500ms between placing MWI calls out the same port. This should allow the CCME enough time to finish processing the last call, and prevent the port from becoming stuck in a SEIZED state. However, this will introduce a slight delay in MWI, if you have a lot of requests queued up. Depending on your hardware and environment, you could get away with setting this value as low as 500ms, but 1500ms is a safe bet across the board.

Release-note: Symptoms: When testing cme with call-gen, it appears that it's possible to exhaust the number of available sockets. We do not know whether this is a normal occurance because the tester tries more then the maximum number of sockets available in IOS unknowingly, or because of a real bug (socket leak). This error message appears: %SKINNYSERVER-3-NOINIT: Can't initialize Skinny server
Conditions: platform confiugred with max number of ephones.
Workaround: Issue is resolved by an architecture enhancement released in 12.4(9)T.

CME/SRST 4.0 when IP phone unregister/ fall back to Call Manager the virtual POTS dial-peers stay up and calls fail with user busy and do not go out the H323 dial-peer to Call manager.

Symptom: Caller ID not displayed on UCC client when using CME 4.3. IP phone tied to UCC does display it.
Conditions: CME 4.3
Workaround: use CME 4.2 or earlier.

Symptom: BUSY tone persists after resuming a transferred call.
Conditions: After the transfer timer expires due to the transfer-to not answering the call and resuming the transferor DN, the busy tone can be still heard.
Workaround: There is no workaround but resetting the phone.

Symptom: Call hung and stuck in active state, so the ephone-dn is stuck in on hook.
Conditions: intermittent. hard to recreate but happens a few times a week.
Workaround: reload GW

Symptom: CME 4.2 does not display calling name when sent in an ISDN facility IE message. The facility is received and interpreted correctly however it doesn't show up on the IP phone display.
Workaround: IOS 12.4(11)XW3 and 12.4(15)XY correct display the calling name.