A fix is available
APAR status
Closed as unreproducible in next release.
Error description
The CPU utilization in the TCP/IP address space is higher in V2R2 than prior releases when sysplex distributing a high volume of short lived connections. The sysplex distribution of a new connection is generating a storage obtain and when the connection terminates there is an associated storage release. Prior to V2R2 the storage object was obtained from a cache queue, thus avoiding the storage obtain and release. The majority of the CPU increase is attributed to contention on the local lock when multiple SRBs are processing concurrently.
Local fix
Problem summary
**************************************************************** * USERS AFFECTED: * * All users of the IBM Communications Server for z/OS Version * * 2 Release 2 IP: Sysplex Distributor and Dynamic VIPA * **************************************************************** * PROBLEM DESCRIPTION: * * Performance slowdown for short lived connections to a DRVIPA * * at the Sysplex Distributor. The slowdown may also be * * present at the target system, or the owning system for a * * Dynamic VIPA. * **************************************************************** * RECOMMENDATION: * * Apply PTF * **************************************************************** Changes introduced in V2R2 caused the Connection Route Hash Table Entry (CRHTE) used to represent a distributed connection to grow in size. The new size did not fit into the storage caching model for TCPIP so all CRHTE control blocks resulted in a storage obtain when the connection initiated and a storage release when the connection terminated. The CRHTE is obtained from TCPIP LSQA which requires the local lock to be held across the storage request. Under heavy loads this additional storage obtain and release may cause contention on the local lock. The contention on the local lock causes units of work to be suspended and resumed which requires additional CPU and introduces delays in the workflow. Targets for Distributed Dynamic VIPAs and owners of Dynamic VIPAs also create a CRHTE to track the state of all Dynamic VIPA related connections in the event the Distrubutor function is moved, or the Dynamic VIPA ownership is moved. The storage obtains and releases for these endpoints may also see an increase in CPU utilization.
Problem conclusion
The storage caching model in TCPIP has been amended to account for the new CRHTE size.
Temporary fix
Comments
APAR Information
APAR number
PI74908
Reported component name
TCP/IP V3 MVS
Reported component ID
5655HAL00
Reported release
220
Status
CLOSED UR1
PE
NoPE
HIPER
YesHIPER
Special Attention
NoSpecatt / Xsystem
Submitted date
2017-01-13
Closed date
2017-02-23
Last modified date
2017-04-04
APAR is sysrouted FROM one or more of the following:
APAR is sysrouted TO one or more of the following:
UI44952
Modules/Macros
EZBITSTO
Fix information
Fixed component name
TCP/IP V3 MVS
Fixed component ID
5655HAL00
Applicable component levels
R220 PSY UI44952
UP17/03/31 P F703 ¢
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":"BU054","label":"Systems w\/TPS"},"Product":{"code":"SG19M","label":"APARs - z\/OS environment"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"220","Edition":"","Line of Business":{"code":"","label":""}},{"Business Unit":{"code":"BU054","label":"Systems w\/TPS"},"Product":{"code":"SSCY4DZ","label":"DO NOT USE"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"220","Edition":"","Line of Business":{"code":"","label":""}}]
Document Information
Modified date:
04 April 2017