What is the parameter session routing for link failure ?

l***@attglobal.net

2008-04-06 14:52:57 UTC

Post by HeungKi-Kim
Hello, :-)
We have following configuration.
=======================================================================
# sna -d l
Link Adjacent Node Device # of local In
station CP name type name State sessions use
-------------- ----------------- ----- --------- ---------- ---------- -----
ETHLINK1 VTAM1.SSCP01 NN ent1 Active 33 Yes
ETHLINK2 VTAM1.SSCP02 NN ent2 Active 33 Yes
# sna -d sl
Max Min Min Act Act Act
Local LU Name Partner LU Name Mode Sess ConW ConL Sess ConW ConL
----------------- ----------------- -------- ----- ---- ---- ----- ---- ----
VTAM1.A0P002 VTAM1.SSCP01 CPSVCMG 2 1 1 2 1 1
VTAM1.A0P002 VTAM1.SSCP02 CPSVCMG 2 1 1 2 1 1
VTAM1.CICSH231 VTAM1.PRODCICS CICSAIX 30 15 15 30 15 15
VTAM1.CICSH231 VTAM1.PRODCICS SNASVCMG 2 1 1 1 0 1
VTAM1.CICSH232 VTAM1.PRODCICS CICSAIX 30 15 15 30 15 15
VTAM1.CICSH232 VTAM1.PRODCICS SNASVCMG 2 1 1 1 0 1
=======================================================================
The LINK fail error happen, and link is reconnected.
Then, sna status is following.
=======================================================================
# sna -d l
Link Adjacent Node Device # of local In
station CP name type name State sessions use
-------------- ----------------- ----- --------- ---------- ---------- -----
ETHLINK1 VTAM1.SSCP01 NN ent1 Active 33 Yes
ETHLINK2 VTAM1.SSCP02 NN ent2 Active 0 Yes
=======================================================================
But,
If we want session routing following status.
=======================================================================
# sna -d l
Link Adjacent Node Device # of local In
station CP name type name State sessions use
-------------- ----------------- ----- --------- ---------- ---------- -----
ETHLINK1 VTAM1.SSCP01 NN ent1 Active 66 Yes
ETHLINK2 VTAM1.SSCP02 NN ent2 Active 0 Yes
=======================================================================
What is the parameter?
thanks...

It is not as simple as one parameter and it
may not be related to any parameters at all.

You appear to have the CS Linux node defined
as a Network Node (because it shows CP-CP
sessions to both VTAMs at the same time)
and the network appears to be using APPN to
locate LUs. In order to use SSCP01 to route
sessions which used to run over SSCP02 you
need to have APPN connectivity between SSCP01
and SSCP02. This is needed so that SSCP02
and SSCP01 can inform each other of which
LUs they have access to, i.e. so that APPN
'locates' can find the target dynamically.

You don't show whether the PRODCICS target
is on a third host or on SSCP01 or SSCP02.
You CANNOT have two "PRODCICS" targets (i.e.
the same partner LU on two remotes at the
same time). LU names must be unique within
an APPN network.

Since you are using APPN, you don't want to
have any define_directory_entry stanzas in CS/AIX
and no ALSLIST= in VTAM. Those are methods of
bypassing the APPN locate (hardcoding locations).

Assuming one of the issues above is not the
problem then the problem could be because
PRODCICS has not tried to reestablish the
sessions to CICSH232 or the application on
AIX has not tried to reestablish the sessions
from CICSH232 to PRODCICS. The sessions have
to be re-started by the application, they won't
be automatically restarted by CS/AIX.

Paul Landay