IMS in the Parallel Sysplex Volume I: Reviewing the IMSplex Technology

198 IMS in the Parallel Sysplex, Volume I: Reviewing the IMSplex Technology

Logon rejected

Then, in Figure 8-15, if the user gets tired of waiting (maybe the user knows that IMS1 will be

down for an extended period of time) and logs on directly to IMS2, IMS2 will check the NodeA

entry and find that NodeA is owned by IMS1 and reject the logon.

Too bad;

however, IMS2 can steal the node if the logon exit (DFSLGNX0) says it is OK. A

clever user shop will write a logon exit to recognize some user data entered at logon as a

request to steal the node. For example, the user data might say, “

I really mean it!”. If this

happens, then the conversation is not recoverable. When IMS1 is restarted and finds the

resource with recoverable status and SRM=LOCAL, IMS1 will check the resource structure

and discover that NodeA/UserA/etc. are no longer owned by IMS1 and delete the status. That

conversation is gone.

Figure 8-15 Status recovery (SRM=LOCAL)

8.8.3 Status recovery (SRM=NONE)

SRM=NONE (or RCVYxxxx=NO) means that end-user status is not to be recovered, even

locally. When a session terminates with SRM=NONE, all end-user status is deleted. See

Figure 8-16.

This might be useful with ETO, which will not delete control blocks if there is any status. For

example, an ETO STSN device ALWAYS has status (always has a sequence number).

SRM=NONE for these devices would delete the status, and then the resources themselves

could be deleted.

VTAM (Generic Resources)

CF

S

C

I

CQS

OM

SCI

IMS

CTL

S

C

I

RM

SCI

SCI

SCI

SCI

IMS2

S

C

I

CQS

OM

SCI

IMS

CTL

S

C

I

RM

SCI

SCI

SCI

SCI

IMS1

LOGON IMS2

NODEA

LTERMA

NODEA

LTERMA

VGR affinity

set to IMS1

SRM=LOCAL

Recoverable

Message

Recoverable Status

Owner=IMS1

SRM=LOCAL