Break Tree Processing

The protected conversation tree (more commonly called the allocation tree) describes the structure of nodes that are allocating protected conversations within a coordinated transaction. See the description of allocation and sync point trees in the z/VM: CMS File Pool Planning, Administration, and Operation. Break tree processing is done to ensure the integrity of the LUWID whenever a protected conversation abnormally ends. An LUWID is associated with each sync point, and it is an SNA LU 6.2 requirement that LUWIDs be unique within the SNA network.

If any error breaks a link in the allocation tree, then two smaller trees are created. An allocation tree can have only one LUWID. The key concept of break tree processing is to let one part of the allocation tree survive and keep the LUWID, and to break apart the other tree and give each piece a different LUWID. If the tree gets broken above you, then you are part of the sub-tree that gets dismantled. If the break occurs below you in the allocation tree, then you are part of the tree that lives.

Figure 1 shows an allocation tree in which VMA talks to VMB, VMB talks to VMC and VMD, and VMD talks to VME.
  • If the protected conversation between VMA and VMB has an error and ends, then break tree processing requires that all the other protected conversations end.
  • If the protected conversation between VMD and VME has an error and ends, then break tree processing requires that no other protected conversations end.
  • If the protected conversation between VMB and VMD has an error and ends, then break tree processing requires that the protected conversation between VMD and VME also ends.
Figure 1. Break Tree Processing
dmsa3cr2

Break tree processing, if it occurs, is part of SPM coordination. The SPM accomplishes the break tree processing by sending an ADADA (Deallocate Abend) action to all resource adapters registered for the work unit. For most resource adapters, this deallocate will be a null operation. If applicable, the resource adapter should make sure that the resource manager is aware that a new LUWID will be used for the next sync point. The new LUWID will not be created by simply incrementing the sequence number field.