How the HADR pair realize that the HADR network is recovered?

When the network between the HADR pair is crashed,
HADR ConnectStatus turns to Disconected after HADR_TIMEOUT.
And then the HADR network is recovered,
the Standby database can realize it and tries to catch up with primary database.

My question is how the HADR pair realize that the HADR network is recovered ?
Are they still sending the keepalive packets for each other after HADR ConnectStatus turns to Disconected?

There is a descirption below on the DB2 info-center,

============================
http://therapy.torolab.ibm.com/infoc...n/c0011748.htm

Standby database states in high availability disaster recovery
7however, if a connection does not exist, the standby
7database tries to connect to the primary database.

============================
From this, standby database seems to keep on checking if he can communicate with primary database by sending keepalive packet.
Does anybody know how it works?

Whenever a running HADR standby is not connected to the configured HADR
primary, it will regularly attempt to establish such a connection over
TCP/IP. This is done without consideration for the current status of the
interconnect and may fail repeatedly until the network is once again
available. Meanwhile, a disconnected HADR primary will constantly listen on
the configured port, waiting for the standby to connect to it.

Regards,
- Steve P.
-----------------------------------------
Steve Pearson
IBM DB2 UDB for LUW Development
Portland, OR, USA

Hello, Steve.
Thank you so much for your help.
The information you give is every time so useful.

Best Regards.

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Tomohiro Iwahashi
IBM Systems Engineering,
Information Management DB2 UDB Technical Support
E-mail ID : bashii-at-jp.ibm.com
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=