Auto resending after a failover

(1) Auto Resending After Failover (Sending Side)

When a failover occurs during sending-side processing (that is, while transferring data), the auto resending can be carried out on the STANDBY node that has taken over the operations. In an event when a failover occurs on the sending side, the auto resending is carried out only if a forced stop was the cause why the sending-side processing instance terminated unsuccessfully.

= Remarks =

  • '590' is the error code of HULFT on the sending side for which the auto resending is carried out after the failover.

  • If the HULFT error code on the sending side is '525,' auto resending is not carried out.

    HULFT on the sending side ends the data transfer and is forcibly stopped in the middle of a termination processing before the Send processing is complete (sending is incomplete).

  • Checkpoint Resend File is not performed.

Figure 1.3 Auto Resending After Failover (Sending Side)

1) Error Detection

The clustering software detects an error (system failure) in the cluster (within the group for failover).

2) Forced stop of the Send process

HULFT on the sending side (ACTIVE node) forcibly stops the Send process.

3) Failover

The STANDBY node takes over the processing from the ACTIVE node.

4) Startup of the Send process

The start synchronization function starts the Send process on the node that has taken over operations.

5) Auto resending after failover

In order to restart the file transfer that was forcibly stopped in 2) , HULFT on the sending side generates a new Send process and connects to HULFT on the receiving side to perform auto resending.

(2) Auto Resending After Failover (Receiving Side)

When a failover occurs during receiving-side processing (that is, while transferring data), the auto resending is carried out on the STANDBY node that has taken over the operation. HULFT on the receiving side notifies HULFT on the sending side that auto resending should be carried out. In an event when a failover occurs on the receiving side, the auto resending is carried out only if a forced stop was the cause why the receiving-side processing instance terminated unsuccessfully.

= Remarks =

'591' is the error code of HULFT on the receiving side for which the auto resending is carried out after the failover.

Figure 1.4 Auto Resending After Failover (Receiving Side)

1) Error Detection

The clustering software detects an error (system failure) in the cluster (within the group for failover).

2) Forced stop of the Receive process

HULFT on the receiving side (ACTIVE node) forcibly stops the Receive process.

3) Notifying HULFT on the sending side

The HULFT on the receiving side notifies the HULFT on the sending side that the Receive process has been forcibly stopped due to an error, and then it disconnects the connection.

4) Failover

The STANDBY node takes over the processing from the ACTIVE node.

5) Startup of the Receive process

The start synchronization function starts the Receive process on the node that has taken over operations.

6) Connection to HULFT on the receiving side

The HULFT on the sending side connects to the HULFT on the receiving side (STANDBY node) after the time set as Connect Retry Interval (retrytime) in the System Environment Settings expires and then transfers the file (auto resend).

7) Auto resending after failover

When auto resending is carried out from HULFT on the sending side, HULFT on the receiving side receives the file after the failover occurs.