How to Reduce the Performance Impact of a Site Recovery
In SCCM, management points can get flooded with policy requests after a disaster recovery. This can especially impact native mode sites. While most mixed- mode customers never notice the impact, native mode uses HTTPS communication, which is much more resource expensive than HTTP communication.
When a site recovery is done, the LastSafeTime value is updated in the database. This will cause all of the client policies to be reset and they will need to be re-signed. Large amounts of client policies can heighten the impact to the management points as many of the clients are contacting the management point with policy requests simultaneously. To lessen the impact, all or some of the below methods below can be employed if using native mode sites. If this is a planned server move or hardware swap, numbers 2-4 should be done before the last backup.
- Restore the site server, but do not deploy any management points.
- Reduce the policy refresh cycle to every 2 days or greater.
- Turn off other items that use management point resources like hardware and software inventory data collection.
- Throttle the number of IIS connections that can come in on the management points.
- Deploy the management points.
- After all clients have refreshed their policies, set the policy refresh cycle and hardware and software inventory configuration back to their previous settings.