IBM Support

Hyper PTF may include longer Abnormal IPL times during C9002C20 SNADS recovery

Flashes (Alerts)


Abstract

Hyper PTFs for SE75867 may include longer Abnormal IPL time during C9002C20 SNADS recovery.

Content

Hyper PTFs for SE75867 - SI76761 (7.4) SI76760 (7.3) SI76759 (7.2) correct a problem where Systems Network Architecture Distribution Services (SNADSrecovery and RCLSTG of SNADS objects would only process the first 1 million SNADS distribution control blocks (DCB) on the system before failing with an MCH0601 for object SNADSTEMPSP.
The remaining DCB objects were not tracked.  They were effectively invisible to the system and could not be cleaned up by normal means.  The specific distributions that were hidden was unpredictable.
Systems with more than a million SNADS DCB objects that have these PTFs applied and performs an abnormal IPL may notice longer SNADS recovery during SRC C9002C20 over similar abnormal IPLs in the past.  The time spent in SNADS recovery due to an abnormal IPL is proportional to the number of DCBs on the system and the system capabilities (memory, processor, DASD subsystem, memory pool sizes).
Note that SNADS recovery is only performed due to an abnormal IPL and will not be performed for a normal IPL.

A diagnostic tool is planned with APAR SE76132 will report the number of SNADS objects, including DCBs, that a system has so that preemptive system SNADS cleanup can be planned. Systems with more than a million DCBs should plan for a maintenance window to run RCLSTG.  This will allow all DCBs on the system to be tracked so that distributions can be cleaned up normally, or force a more aggressive cleanup activity to remove the excess DCB objects.
As an alternative for the planned diagnostic tool, a customer could use the new Advanced Analysis macro: "AUINTERNALS" which offers a way for customers to check whether they will have an impact on their next Abnormal IPL or SNADS recovery in specific.
The number of objects owned by the QSNADS user profile can provide some idea if the system will be impacted.
See the attached document related on that.
Cleanup activities should include using QMGTOOLS/CLEANMAIL for SDD users that have old distributions
NOTE: CLEANMAIL cannot be used to remove any QNONDELI distributions created by SMTP for Non-delivery notifications if configured,
or any SDD entries that start with 'Q' due to the restriction of submitting jobs for 'Q' system users.
For more information on QNONDELI distributions: "Creating a Non-delivery mailbox"
For more information on QMGTOOLS or how to obtain is see the attached link: "Mustgather How to obtain and Install QMGTOOLS and keep it current".
Distributions to SNA nodes that no longer exist should also be cleaned up using INZDSTQ and/or removing the obsolete distribution queues.
Other internal system limits occur around 31 million objects which will exceed the size that the SNADS distribution index (QZDINDEX) can contain. Systems with more than 31 million DCBs are most likely the result of a system with a rogue SNADS/Mail application that has flooded the system. 
A final limit exists of 50 million objects which exceeds the total number of objects that can be owned by the QSNADS user profile.
SNADS objects owned by QSNADS include DCBs, FSOs, DTOs, DUOs and other user visible object types.
IBM Service should be consulted on support options to cleanup from extreme numbers of SNADS objects.

[{"Type":"MASTER","Line of Business":{"code":"LOB57","label":"Power"},"Business Unit":{"code":"BU058","label":"IBM Infrastructure w\/TPS"},"Product":{"code":"SWG60","label":"IBM i"},"ARM Category":[{"code":"a8m0z0000000CLNAA2","label":"Communications->SNADS"},{"code":"a8m0z0000000CIjAAM","label":"Email and SMTP->SNDDST"}],"ARM Case Number":"","Platform":[{"code":"PF012","label":"IBM i"}],"Version":"7.2.0;7.3.0;7.4.0"}]

Document Information

Modified date:
24 February 2022

UID

ibm16484665