Showing results for 
Search instead for 
Did you mean: 


Scheduled Periodic Rebooting of CM/Unity/IPIVR (CRA)

After performing troubleshooting of various IPT implementations for different clients, I have experienced numerous phone, voice mail, and CTI issues that were resolved by performing simple restarts (usually within CallManager) on several components, including Attendant Console Pilot Points, Unity VM Ports, and CTI Route Points. I'm sure this isn't news to most of you. In most cases, these types of issues cropped up after the servers had been running for a few months since the last server reboot.

I'm considering recommending to one of my clients that they perform periodic rebooting of their IPT servers with the intent of minimizing these miscellaneous issues. As I have noticed on the forums, scheduling Unity reboots appears to be quite common.

The versions that we are using are as follows:

CM 3.3(3) sr3

Unity 4.0(3)

CRA 3.1(2) - IPIVR

This particular implementation has around 200 IP Phones/mailboxes and 16 unity ports, and will likely grow to max 750 users per CM/Unity Cluster.

So my questions are:

(1) Outside of keeping the servers patched on a regular basis, how often do you recommend performing periodic reboots of IPT servers? Daily? Weekly? Monthly?

(2) Do any of you NOT recommend periodic rebooting of servers? If so, why?

I realize that there could be many factors, such as number of phones, scripting configuration, utilization, server platform, Operating system, etc,. Just a general idea would be helpful. Thanks.


Reboots have helped many of my customers, solve with Unity voice mail and CallManager. With older versions of Unity, I have seen voice mail stop responding, many times, which could be solved only by rebooting the machine..(the only other fix Cisco could suggest was an upgrade to higher version). I have recommended this customer, to do a scheduled reboot of servers, once a month..Ofcourse they do have a schedule reboot for many of the other servers, so this was just an addition to the list...

I dont see a reason why one should not do a periodic rebooting of long as you have multiple servers hosting the phones..

Sankar Nair
UC Solutions Architect
Pacific Northwest | CDW
CCIE Collaboration #17135 Emeritus
Daniel Baum
Cisco Employee

For Unity: the specific answer(s) to your question(s) is here:

Restarting the Cisco Unity Server

Depending on how your network is configured, Cisco Unity services may rely on the availability of a number of other servers on the network, as follows:

The Exchange server that homes the Unity System mailbox The Exchange server(s) that home Cisco Unity subscriber mailboxes Exchange bridgehead servers

The domain controller for the domain that the Cisco Unity server belongs to The Active Directory global catalog server The Cisco CallManager server (if Cisco Unity is integrated with Cisco CallManager)

The Cisco Unity secondary server, for Cisco Unity failover systems The availability of each of these servers, the network connection to these servers, and the corresponding components on the Cisco Unity server may each affect Cisco Unity services.

If you restart servers or network devices on a regular schedule, include Cisco Unity servers in this schedule. See the "Exiting and Starting the Cisco Unity Software and Server" section.

For CallManager: Unless a specific bug is being hit that requires a reboot, there is not a published maintenance routine for reboots, etc.

Just my opinion, if perhaps the CallManager OS is a part of a Windows domain or Directory Integrated with AD, etc., and borrowing words from the Unity doc, "depending on how your network is configured, Cisco CallManager OPERATING SYSTEM services may rely on the availability of a number of other servers on the network. a reboot may be in order from time to time. Otherwise, if CallManager is running independant or not, and no obvious errors or issues, let it run.



Hello David,

I tell my customers to reboot their servers once a week. I have had an issue once in a while with Unity not starting up as it should, but in most cases setting the services recovery to 'restart the service' has done the trick. I try and schedule the reboot so that the customers can check the servers out before business operations resume.

Recognize Your Peers
Content for Community-Ad