Configuring Graceful Timeout
Many maintenance actions on Media Components (such as locking, restarting or upgrading) let you employ a "grace" period (graceful timeout) before the action is started by the Cluster Manager. The graceful timeout lets the Media Component continue processing currently active media sessions until the timeout expires. During this timeout, the Cluster Manager doesn't allocate new media sessions to the Media Component. If all media sessions end before the timeout expires, the Cluster Manager doesn't wait for the timeout to expire, but performs the maintenance action immediately.
If active media sessions still exist when this timeout expires, the Cluster Manager tries to transfer the active sessions (sending re-INVITE messages) from the Media Component (e.g., MC #1) on which it needs to perform the maintenance action, to other Media Components, and then performs the required maintenance action on the Media Component (e.g., MC #1). Sessions that cannot be transferred due to unavailable resources on the other Media Components are immediately terminated. These terminated sessions are indicated in the syslog messages.
|
➢
|
To configure the graceful timeout: |
|
1.
|
Open the Cluster Manager Settings page (Setup menu > IP Network tab > Media Cluster folder > Cluster Manager Settings). |
|
2.
|
In the 'Cluster Graceful Timeout' field, enter the graceful timeout: |
