06-25-2010 10:55 AM - edited 03-19-2019 01:09 AM
Clustered Unity 7.12, local message store with two phone integrations (Callmanager and Centrex) with ~6400 users. I have heard from two users this week that they have experienced delayed voice mail message delivery, seemed to be delivered next day. The issue appeared on both phone integrations. The user's quota looked fine. Any suggestions?
Thanks,
Beth
06-25-2010 11:06 AM
I believe you mean Unity Connection, right? If so, take a look here for starters: http://www.cisco.com/en/US/docs/voice_ip_comm/connection/7x/troubleshooting/guide/7xcuctsg120.html#wp1051293
Hailey
Please rate helpful posts!
07-22-2010 10:50 AM
Thanks David however the report(s) did not give me any real information. Maybe I did not generate the correct report, I gerated a User Report and a Message Traffic Report. I got another report of a delayed message this morning ( 1/2 hour) and again the reports did not provide any useable information. Any other ideas. thanks, Beth
07-22-2010 01:31 PM
What is your setup? Do you have a Unity Connection cluster or standalone box? Integration with Centrex is via PIMG? Where are users located in relationship to your cluster? Have you considered upgrading to a more stable version of code - such as 7.1(3b)SU2? 7.0 and 7.1(2) did not lack in the "having bugs" department so it's more of a question than a suggestion of cause.
Hailey
07-22-2010 02:26 PM
We are a cluster, users and servers distributed across a campus. We have two phone integrations, VoIP and Centrex with PIMGS. The users experiencing the delays are on the VoIP integration. We are planning to upgrade in August to some flavor of 7.1.5.
Thanks,
Beth
07-22-2010 02:38 PM
Since the users are experiencing delays on the VoIP integration, I do have a couple questions:
How many ports in the VoIP integration are dedicated to answering calls only?
How many ports in the VoIP integration are dedicated to MWI/Callout (all functions EXCEPT Answering Calls)?
How do you have your line groups and hunt lists configured in CUCM? In other words, the Subscriber ports (minus MWI ports) should be preferred and included in a separate line group that is a priority in the hunt list. The Publisher ports (minus MWI) should be included in a separate line group and listed as secondary in the hunt list. Is this the case?
Are the users that are experiencing delays strictly doing voicemail and/or Integrated Messaging? In other words, this is not a case where users are having messages forwarded via SMTP relay to a non-Unity Connection SMTP address such as a corporate email, correct?
Lastly, the delay appears is manifested in MWI On/Off notification only? In other words, users gets notification of a message but the timestamp of the message indicates it was left x minutes (or some period of time) earlier than when the MWI was received?
Hailey
07-22-2010 02:38 PM
Sorry, one more - is the VoIP integration with CUCM? Is it a SCCP or SIP integration?
07-23-2010 10:45 AM
Hailey,
We will be upgrading to 7.13 in the next few weeks.
Here are the answers to your inquiries, thanks.
Beth
How many ports in the VoIP integration are dedicated to answering calls only?
96 ports
How many ports in the VoIP integration are dedicated to MWI/Callout (all functions EXCEPT Answering Calls)?
None, all 96 VoIP integrated ports are set for Answer Calls, Message Notification and Dialout MWI.
How do you have your line groups and hunt lists configured in CUCM? In other words, the Subscriber ports (minus MWI ports) should be preferred and included in a separate line group that is a priority in the hunt list. The Publisher ports (minus MWI) should be included in a separate line group and listed as secondary in the hunt list. Is this the case?
No.
Are the users that are experiencing delays strictly doing voicemail and/or Integrated Messaging? In other words, this is not a case where users are having messages forwarded via SMTP relay to a non-Unity Connection SMTP address such as a corporate email, correct?
Users are just doing voice mail.
Lastly, the delay appears is manifested in MWI On/Off notification only? In other words, users gets notification of a message but the timestamp of the message indicates it was left x minutes (or some period of time) earlier than when the MWI was received?
You know I am not sure if it is just the MWI or the message itself is delayed. The 3 users who experienced the issue are out today. I do know that the MWI is definitely delayed and the actual message has a timestamp of when the message was sent by the caller. I was under the impression that one user checked their voicemail and had no messages, when she checked the next day she had a message dated the day before.
Sorry, one more - is the VoIP integration with CUCM? Is it a SCCP or SIP integration?
SCCP Integration.
07-23-2010 12:08 PM
Ok. Well, typically it's best practice to slice things up a bit. With Unity Connection, the Subscriber should be handling the majority of all calls on a regular basis. I go with a 75% rule. So, let's say I have 48 ports per server - 75% of those are answering ports only. So, on each server - you'd have 36 ports that answer calls and 12 ports that are MWI, Message Notification, etc (but do not answer calls).
From there, the Subscriber ports (answering ports only) are configured in an individual line group.
Likewise, the Publisher ports (answering ports only) are configured in an individual line group.
Then the VM hunt list would have the Subscriber line group listed first and the Publisher listed second. You should also hunt top down and make sure that you enable rolling over to the next line group.
In CUC, you'd have 2 port groups. 1 for the Sub and 1 for the Pub. You create your answering ports for each line group first (and only select answer calls) and then the MWI/dialout ports for each group second.
Keep an eye on what happens when you move to 7.1(3). If you continue to have problems, you may want to reconsider the port configuration to see if that helps.
Hailey
07-23-2010 11:30 PM
I just wanted to tack on the recommendation that I think you seriously need to reconsider your port configuration. When dealing with a cluster, it is very important that you set things up based on the supported guidelines provided by Cisco. For example:
Under normal conditions, the Cisco Unity Connection cluster will handle calls in the following manner:
•The subscriber server will answer most incoming calls. If no answering ports on the subscriber server are available, the publisher server will answer calls.
•The publisher server will dial out for MWIs and notifications If no dial-out ports on the publisher server are available, the subscriber server will dial out for MWIs and notifications.
In your scenario, if the ports are lumped together and all ports do everything then it would be possible for the servers to get delayed from time to time simply based on configuration. Read thru this guide which will confirm what I've told you (the only thing that is not specifically stated is the 75% answering ports / 25% other functions - MWI, etc. port usage ratio. This is a general recommendation many engineers use and I've found it works well for all of my clusters - some of which are very large, ex: 2 clusters, networked, 10000 users and growing, etc).
This is the section on clustering. The general config guidelines are here:
Hailey
Please rate helpful posts!
Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: