DEFINITY Enterprise Communications Server Release 7
Maintenance for R7r
555-230-126
Issue 4
June 1999
Additional Maintenance Procedures
6-3Software updates
6
If these conditions are not met, The
upgrade software call-override
option must
be used, and the update will result in a system reset level 2 (cold-2 restart). An
additional option,
preserve-calls
, can be added to force the system to attempt to
preserve calls, but this must be used with care. If the incompatibility is great
enough, this option may cause the system to escalate to a reboot (
reset system
4
). It is recommended that you escalate any such software update.
Call-Preserving Update
During a call-preserving update, calls in which two or more parties are
connected and talking are preserved. The following types of calls are dropped:
wideband calls, held calls, dialing calls, and calls that are connected to
announcements, speech synthesizers, or tones. Administered connections are
dropped and automatically restored. Feature activation attempts are ignored. No
new calls are processed for a brief period. This period usually lasts less than 5
seconds but can last up to 1 minute depending on traffic load during the SPE
interchange. Conference calls that attempt a new connection are either ignored
or dropped.
Call-Dropping Update
During a call-dropping update, the effect will be similar to a cold-2 restart (
reset
system 2
). All calls and system links drop, and the MT login is terminated. The
service outage lasts up to 4 minutes.
Preparing for a Software Update
1. Notify users of anticipated service effects described in the preceding
section and arrange to do the update at a suitable time. Also advise
appropriate users of the following feature interactions:
■ Administered connections are temporarily dropped.
■ Leave Word Calling messages stored in the SPE are lost. Those
stored in a Message Server Adjunct or AUDIX system are saved.
■ Some feature settings will be lost and must be restored afterward.
The effect is the same as a cold-2 restart. Features affected include
night service, trunk/hunt group control, and manual/clock-manual
override status on time of day routing. For example, hunt groups
and attendant are put into night service after the update.
■ ACD queues will be lost. There is no mechanism to drain the
queues. CMS or BCMS can be used to monitor the queue lengths.
If it is important to empty all queues before the update, you must
coordinate with the customer to redirect traffic away from ACD
splits before the update. Changing vectors to temporarily route
calls to a disconnect announcement is one way to accomplish this.
ACD agents will need to log in again after the update and their
status lamps may be incorrect for about an hour (until periodic
background maintenance completes one cycle).
