Deciding How to Restore Service
Topic Last Modified: 2009-02-28
To minimize the impact on users, restoration of service should be done in a way that causes the least disruption while still bringing the environment back to an acceptable level of service. Your backup and restoration plan should contain criteria for deciding when and how to restore service. In the event of loss of service, use the criteria to determine what is most appropriate for resolving the problem. For example, the criteria should help you decide whether you only need to recover data and/or settings or whether more extensive server restoration is required for one or more servers, pools, or sites.
Note
To reduce down time and potential loss of data, you should perform the restoration procedures described in this document only if troubleshooting procedures are not effective in identifying and correcting a problem with any individual server or component. Shutting down and restarting servers during troubleshooting should be done in a way that minimizes the impact on other components and servers.
If you need to restore multiple servers, you must restore them in the appropriate sequence. The following table indicates the restoration sequence for each type of deployment. Use the sequence shown, skipping any servers, such as Edge Servers, that are not in your deployment.
Table 1. Restoration Sequence for Servers
| Servers | Standard Edition server environment | Enterprise pool, consolidated configuration |
|---|---|---|
Back-end servers |
Not applicable |
1 |
Standard Edition servers |
1 |
Not applicable |
Front End Servers |
Not applicable |
2 |
Directors |
Not applicable |
3 |
Archiving Servers |
2 |
4 |
Group Chat Servers and Group Chat Compliance Servers |
3 |
5 |
Monitoring Servers |
4 |
6 |
Communicator Web Access Servers |
5 |
7 |
Mediation Servers |
6 |
8 |
Edge Servers |
7 |
9 |
Forwarding proxy servers |
8 |
10 |
Reassigning users |
9 |
11 |
Restoring domain information |
10 |
12 |
Use the information in this topic, the procedures in this guide, and the information available in your organization to complete the restoration of service for each database, server, or site that requires restoration. In general, the procedures that you use and the sequence in which you complete them should be determined by which components need to be restored (such as a site, servers, and databases), as described in Figure 1.
Figure 1. Restoration Processes
.jpg "Dd572205.2e8b5142-229b-4607-a029-3ebb6c36d80d(en-us,office.13).jpg")
The following table contains links to the sections referenced in the preceding flowchart.
Restoring Settings in Enterprise Pool, Consolidated Topology Environments |
Group Chat has the following component-specific restoration guidelines in the event of loss of service for the following components:
- Channel Servers:
- If service can be restored using the existing Channel Servers, no additional action is required.
- If service cannot be restored using the existing Channel Servers, restore service using new Channel Servers.
- Group Chat database:
- If service can be restored using the existing Group Chat database, no additional action is required.
- If service cannot be restored using the existing Group Chat database, restore service using a new database and restoring the data from the backup location, configure the Channel Servers and Compliance Server to point to the new database, and then restart the servers.
- Compliance Server:
- If service can be restored using the existing Compliance Server, no additional action is required.
- If service cannot be restored using the existing Compliance Server, restore service using a new server, configure the Compliance Server to point to the new database, and then restart the Channel Servers. Any messages in the old Compliance Server Message Queue will be lost.
- Compliance database:
- If service can be restored using the existing compliance database, no additional action is required.
- If service cannot be restored using the existing compliance database, restore service using a new database, configure the Compliance Server to point to the new database, and then restart the Compliance Server. Any messages in the old Compliance Server Message Queue will be lost.