Understanding NSPI Connection Limit Reached

In some deployments, the ArchiveOne Service can run out of NSPI connections. This is due to a change in the operation of Microsoft's RPC layer code that is used when the ArchiveOne Service is connected to a Domain Controller to gather information about mailboxes. This change causes resources to be used up and incorrectly released. The resources eventually reach the limit, causing the RPC connections to fail.

If you encounter one or more of the following issues:

• Policies fail every few days and ArchiveOne displays the Unable to connect to Exchange error message until the service is restarted
• Retrievals begin failing for all users until the service is restarted
• Administration console connection to the ArchiveOne Server fails until the service is restarted

To reduce the frequency of this problem,

1. Raise the default number of allowed NSPI connections as described in the Microsoft Support NSPI Connection article.

   Barracuda Networks recommends raising default value to 500, however, you should determine a value that best suits your environment without negatively affecting other infrastructure and applications. You can set a simple scheduled task to execute: net stop AOnePolService && net start AOnePolService

2. Restart the ArchiveOne Service at regular intervals.

   Barracuda Networks recommends restarting the service every two to three days, however, you should determine a schedule that best suits your organization, taking into account the policies that are run as well as the time and schedule of those policies.