If a file transfer is disrupted due to a network problem or limited server capacity, syncdat may pause for a while then try to repeat the pending transaction. This behavior is controlled by the TryInterval or -t option.
The default is to try again at one minute intervals until the transaction succeeds or a fatal error occurs.
The TryInterval or -t option may be set to one of the following values:
|fail||The entire session will be terminated. This setting is recommended for scripted scenarios where a parent process may wish to handle recovery itself.|
|skip||The transaction which generated the error will be skipped. If there are other transactions pending, syncdat will begin the next one. This not generally recommended.|
|0||The failed transaction will be retried at the default interval until the transaction succeeds or a fatal error occurs. This is the default.|
|<seconds>||The failed transaction will be retried at the given interval until the transaction succeeds or a fatal error occurs. For connectivity errors, a minimum interval of 60 seconds will be enforced to ensure that the network and server have time to recover. Capacity errors will honor lower intervals.|
The time interval is measured from failure to retry. If the network is completely down, the time measured from retry to retry may be as much as TryInterval plus Expire. You can combine TryInterval with the StopTime or -S option to limit how long syncdat will keep trying.
If the target is a host group, syncdat will search for an available server amongst the group. This may result in some actions being performed through different servers. It is vital that all servers in a host group have identical storage, and that the first server of the group (the Key Server) does not change between runs.
If the host group includes fallback groups, the fallback servers will be used only if all servers in the earlier peer groups are completely unreachable. If servers in a peer group are reachable but at full capacity, syncdat will retry within that group until capacity becomes available.
If the network or server is disrupted during an upload, the server will enforce a mandatory 60 second cooling-off period before that upload can be resumed. Restarting that upload prior to the end of this period may fail with a message indicating that the file may be in use. This will normally result in another automatic retry period.