Ibm Data Server Driver For Jdbc And Sqlj Support For Client Affinities For Cascaded Failover - IBM DB2 Manual

|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
IBM Data Server Driver for JDBC and SQLJ support for client
|
affinities for cascaded failover
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1. The IBM Data Server Driver for JDBC and SQLJ tries to connect to the
Connection Manager that is identified by host1:port1.
2. The connection to host1:port1 fails, so the driver tries to connect to the
Connection Manager that is identified by host2:port2.
3. The connection to host2:port2 succeeds.
4. The driver retrieves alternate server information from the connection manager
that is identified by host2:port2, and updates its memory with that
information. Assume that the Connection Manager identifies host1a:port1a as
the new primary server, and host2a:port2a as the new alternate server.
5. The driver tries to connect to host1a:port1a.
6. The connection to host1a:port1a fails.
7. The driver tries to connect to host2a:port2a.
8. The connection to host2a:port2a fails.
9. The driver waits two seconds.
10. The driver tries to connect to host1a:port1a.
11. The connection to host1a:port1a fails.
12. The driver tries to connect to host2a:port2a.
13. The connection to host2a:port2a fails.
14. The driver waits two seconds.
15. The driver tries to connect to host1a:port1a.
16. The connection to host1a:port1a fails.
17. The driver tries to connect to host2a:port2a.
18. The connection to host2a:port2a fails.
19. The driver waits two seconds.
20. The driver throws an SQLException with error code -4499.
Client affinities for cascaded failover is supported for IBM Data Server Driver for
JDBC and SQLJ type 4 connectivity.
Cascaded failover occurs during client reroute, when primary database servers,
primary standby database servers, and secondary standby database servers are set
up. When the primary server fails, connections are routed to the primary standby
server. When the primary standby server fails, connections are routed to the
secondary standby server.
In a DB2 for z/OS data sharing environment, primary and standby servers
correspond to data sharing group members. The default behavior for cascaded
failover is that the connected data sharing member is the primary server, and other
members in the group are alternates. The connected member is at the top of the
alternate server list that is returned from the connected server, and the other
members are listed in order of capacity. Suppose that there are three members in
an alternate server list: a primary server and two potential failover members.
When the primary member fails, the failover member with the highest capacity
becomes the primary server, and the other failover member becomes the failover
server. Because the order is determined by capacity, the order can change. If you
want to enforce a specific failover order for certain applications, you can enable
support for client affinities for cascaded failover.
Chapter 3. JDBC application programming
95