Occasionally, a full deployment can stop in an incomplete state due to a non-fatal error—for example, an SQL deadlock or a connection timeout. In cases like these, you can manually resume the deployment through the Business Control Center Admin Console. Alternatively, you can automate the recovery process by setting the DeploymentManager property transientErrorRetryCount. This property sets the number of times a deployment thread tries to redeploy a transaction batch after a recoverable error. This property’s default value is 0, which disallows retries.

transientErrorRetryCount is primarily aimed at facilitating recovery from SQL deadlock errors, though not exclusively. Deadlocks are data dependent and some databases handle them better than others. You should set this property to a reasonable value for your environment—typically, between 1 and 5.