PingDataSync is designed to quickly synchronize data and attempt a retry should an operation fail for any reason. The retry mechanism involves two possible retry levels, which are configurable on the Sync Pipe configuration using advanced Sync Pipe properties. For detailed information, see the PingDataSync Reference Guide for the Sync Pipe configuration parameters.

Retry involves two possible levels:

First Level Retry
If an operation fails to synchronize, the server will attempt a configurable number of retries. The total number of retry attempts is set in the max-operation-attempts property on the Sync Pipe. The property indicates how many times a worker thread should retry the operation before putting the operation into the second level of retry, or failing the operation altogether.
Second Level Retry
Once the max-operation-attempts property has been exceeded, the retry is sent to the second level, called the delayed-retry queue. The delayed-retry queue uses two advanced Sync Pipe properties to determine the number of times an operation should be retried in the background after a specified delay.

Operations that make it to this level will be retried after the failed-op-background-retrydelay property (default: 1 minute). Next, PingDataSync checks the max-failedop-background-retries property to determine the number of times a failed operation should be retried in the background. By default, this property is set to 0, which indicates that no background retry should be attempted, and that the operation should be logged as failed.


Background operations can hold up processing other changes, since PingDataSync will only process up to the next 5000 changes while waiting for a retried operation to complete.

Retry can be controlled by the custom endpoint based on the type of error exception. When throwing an exception, the endpoint code can signal that a change should be aborted, retried a limited number of times, or retried an unlimited number of times. Some errors, such as endpoint server down, should be retried indefinitely.

If the max-failed-op-background-retries property has been exceeded, the retry is logged as a failure and appears in the sync and the sync-failed-ops logs.