PDT ECSA cell pool growth for inactive tasks
Article ID: 124198
Updated On:
Products
Bind Analyzer for DB2 for z/OS
SQL-Ease for DB2 for z/OS
SYSVIEW Performance Management Option for DB2 for z/OS
Plan Analyzer for DB2 for z/OS
Subsystem Analyzer for DB2 for z/OS
Issue/Introduction
ECSA storage problemes, so we had a closer look, subpool 241 key 7 looks to be the causer.
We are running in to ECSA storage problemes.
Analyzing ECSA storage shows subpool 241 key 7 as a causer, storage header always CPOOL CELL POOL.
Shows that these ECSA CELL storage pools are from jobs, STCs which do not run any more, so those storage is left as "carcasses" and fills up ECSA.
Looking at CADB2 documentation and lots of hints found, showing Detector as a causer.
E.g
TITLE: DETECTOR ECSA ITH$ENTY CELL POOL GROWTH FROM CPU PARALLELISM but these issues were prior to R20.
We are running in to ECSA storage problemes.
Analyzing ECSA storage shows subpool 241 key 7 as a causer, storage header always CPOOL CELL POOL.
Shows that these ECSA CELL storage pools are from jobs, STCs which do not run any more, so those storage is left as "carcasses" and fills up ECSA.
Looking at CADB2 documentation and lots of hints found, showing Detector as a causer.
E.g
TITLE: DETECTOR ECSA ITH$ENTY CELL POOL GROWTH FROM CPU PARALLELISM but these issues were prior to R20.
Environment
z/os DB2
Resolution
Cell Pools for ITH$ENTY will be reused.
Once ITH$ENTY are allocated, they’ll not be freed until the Detector collection terminates and/or
all SQL statement tracked by Detector has been processed completely.
Those control blocks are allocated and reused for performance reason.
CSA storage for ITH$ENTYS is used for tracking SQL activity for Detector collections and are acquired as needed
and will normally not exceed 500 separate ITH$ENTYs per collection.
This storage is reused as needed and will be released when the collection is terminated.
There are however times when the collection is terminating but there remains in-flight
SQL making it necessary to defer the cleanup of ITH$ENTYS in storage. When this occurs,
a message is issued in the Xmanager log to indicate that storage cleanup is deferred and will
be cleaned up on the next Detector collection startup.
the message in the xmanager log will look like as shown below.
PDT0736 DETECTOR DEFERRED CLEANUP INITIATED DB2=XXXX ID=XXXXX
PDT0737 DETECTOR DEFERRED CLEANUP ENDED DB2=XXXX ID=XXXX
Once ITH$ENTY are allocated, they’ll not be freed until the Detector collection terminates and/or
all SQL statement tracked by Detector has been processed completely.
Those control blocks are allocated and reused for performance reason.
CSA storage for ITH$ENTYS is used for tracking SQL activity for Detector collections and are acquired as needed
and will normally not exceed 500 separate ITH$ENTYs per collection.
This storage is reused as needed and will be released when the collection is terminated.
There are however times when the collection is terminating but there remains in-flight
SQL making it necessary to defer the cleanup of ITH$ENTYS in storage. When this occurs,
a message is issued in the Xmanager log to indicate that storage cleanup is deferred and will
be cleaned up on the next Detector collection startup.
the message in the xmanager log will look like as shown below.
PDT0736 DETECTOR DEFERRED CLEANUP INITIATED DB2=XXXX ID=XXXXX
PDT0737 DETECTOR DEFERRED CLEANUP ENDED DB2=XXXX ID=XXXX
Feedback
Yes
No
Powered by
