RP/MSQ 4.0A(Unix/NT) - RP45 reconnect fails with mulithomed systems if downtime is greater than 2 minutes. (Doc ID 768823.1)

Last updated on JANUARY 16, 2023

Applies to:

Goal

In an XGROUP which has multiple IP addresses, suppose a host which is set Initiate=Y loses its connection.  If this
disconnected condition lasts for more than 2 minutes, then the host will not try to reconnect any longer.  This
appears to be a bug introduced in RP45 for MessageQ V4.0A for Digital UNIX.

Host <HOST1>
G023            23  <JOB_ID> Y  2000000    -1  5  5 25000 TCPIP 30000
G023            23  <JOB_ID> Y  2000000    -1  5  5 25000 TCPIP 30000
G024            24  <JOB_ID> Y  2000000    -1  5  5 25000 TCPIP 30000
G024            24  <JOB_ID> Y  2000000    -1  5  5 25000 TCPIP 30000

Host <HOST2>
G023            23  <JOB_ID> N  2000000    -1  5  5 25000 TCPIP 30000
G023            23  <JOB_ID> N  2000000    -1  5  5 25000 TCPIP 30000
G024            24  <JOB_ID> Y  2000000    -1  5  5 25000 TCPIP 30000
G024            24  <JOB_ID> Y  2000000    -1  5  5 25000 TCPIP 30000

A connection is established between <HOST1> and <HOST2>.  If it is lost by e.g. shutdown of dmq on<HOST2>,
then <HOST1> sends reconnection packets to<HOST2>, because <HOST1> is set Initiate=Y.  Problem is that this
reconnection attempt lasts for only 2 minutes.  If <HOST2>'s dmq gets back up after a longer downtime than 2 minutes,
the connection cannot be re-established.  We see this problem on both es0 and es1.

RP42  and 44 don't appear to have tjhis problem.  This also  works fine with an XGROUP with only one IP address.

Solution

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.