A fix is available
APAR status
Closed as program error.
Error description
Using the WLMAWORK view, and hyperlinking to the AORs (WLMAWAOR) you try QUIESCING the workload in the AOR's with the SET/QUIESCE COMMAND. EYUEI0558E is issued, indicating that an internal error occurred in WAUO. The WLMAWAOR screen shows the AOR's in the quiescing state forever. The AUXTRACE from the CMAS shows an exception trace entry coming out of XSCG with DEBUG:GETMLE (LENGTH ERROR). CPLE had requested storage based on the number of CMAS's in the SETDIR, and the SETDIR_CMASC_CNT was X'FFFFFFFF', so the storage calculation was wrong.
Local fix
Problem summary
**************************************************************** * USERS AFFECTED: All CICSPlex/SM V3R1M0 Users * **************************************************************** * PROBLEM DESCRIPTION: When two or more CMASes join a workload * * (EYULOG message EYUWM0400I) at exactly * * the same time, errors may occur when * * subsequent activities are performed * * against any workload in the CMAS * * network. * **************************************************************** * RECOMMENDATION: After applying the PTF that resolves this * * APAR, all CMASes must be restarted. Note * * that the restarts do not need to occur at * * the same time. * **************************************************************** When a CMAS joins a workload, method EYU0CPAE (CPAE) runs in all CMASes in the CMAS network, updating the workload SET to include the CMAS. When a CMAS exits a workload, method EYU0CPDE (CPDE) runs in all CMASes in the CMAS network, updating the workload SET to remove the CMAS. When an update is made to the workload (e.g., quiesce of an AOR, build of a global affinity, removal of an active WLMDEF, ...), the local CMAS uses the SET to inform all CMASes participating in the management of the workload of the change. Part of the update performed by CPAE is to increment the count of CMASes in the SET. To do this, CPAE retrieves the current count, increments it by one, and then stores the updated count. The process of storing the updated count is serialized by CPAE acquiring the network lock. However, the lock is acquired after CPAE retrieves the current count. This can result in two or more concurrent tasks running CPAE using the same value for the current count. This will cause the updated current count to be less than what it should be. If the CMAS count is lower than it should be, the local CMAS may not notify all CMASes of a change to a workload; this will cause an update to not be performed completely. Additionally, when all CMASes exit the workload, CPDE will decrement the CMAS count for the SET to a negative value; this may cause failures when subsequent update requests are made for other workloads.
Problem conclusion
CPAE has been updated to acquire the network lock prior to retrieving the workload SET CMAS count and will continue to hold the lock until updating of the set is complete.
Temporary fix
********* * HIPER * ********* FIX AVAILABLE BY PTF ONLY
Comments
When two or more CMASes join a workload (EYULOG message EYUWM0400I) at exactly the same time, subsequent errors may occur when activities are performed against any workload in the CMAS network. When a CMAS joins a workload, method EYU0CPAE (CPAE) runs in all CMASes in the CMAS network, updating the workload SET to include the CMAS. When a CMAS exits a workload, method EYU0CPDE (CPDE) runs in all CMASes in the CMAS network, updating the workload SET to remove the CMAS. When an update is made to the workload (e.g., quiesce of an AOR, build of a global affinity, removal of an active WLMDEF, ...), the local CMAS uses the SET to inform all CMASes participating in the management of the workload of the change. Part of the update performed by CPAE is to increments the count of CMASes in the SET. To do this, CPAE retrieves the current count, increments it by one, and then stores the updated count. The process of storing the updated count is serialized by CPAE acquiring the network lock. However, the lock is acquired after CPAE retrieves the current count. This can result in two or more concurrent tasks running CPAE using the save value for the current count. This will cause the updated current count to be less than what it should be. If the CMAS count is lower than it should be, the local CMAS may not notify all CMASes of a change to a workload; this will cause an update to not be performed completely. Additionally, when all CMASes exit the workload, CPDE will decrement the CMAS count for the SET to a negative value; this may cause failures when an update request is made for other workloads. CPAE has been updated to acquire the network lock prior to retrieving the workload SET CMAS count and will continue to hold the lock until updating of the set is complete.
APAR Information
APAR number
PK20038
Reported component name
CPSM CICS 3.1
Reported component ID
5655M1501
Reported release
100
Status
CLOSED PER
PE
NoPE
HIPER
YesHIPER
Special Attention
NoSpecatt
Submitted date
2006-02-17
Closed date
2006-03-20
Last modified date
2006-06-05
APAR is sysrouted FROM one or more of the following:
APAR is sysrouted TO one or more of the following:
PK20108 UK12778 PK25862
Modules/Macros
EYU0CPAE
Fix information
Fixed component name
CPSM CICS 3.1
Fixed component ID
5655M1501
Applicable component levels
R100 PSY UK12778
UP06/03/22 P F603
Fix is available
Select the PTF appropriate for your component level. You will be required to sign in. Distribution on physical media is not available in all countries.
[{"Business Unit":{"code":"BU058","label":"IBM Infrastructure w\/TPS"},"Product":{"code":"SSGMGV","label":"CICS Transaction Server"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"3.1","Edition":"","Line of Business":{"code":"LOB35","label":"Mainframe SW"}}]
Document Information
Modified date:
22 February 2023