PH13377: SOME TABLE OPERATIONS INVALIDATE AND RELOAD THE BIGSQL SCHEDULER CACHE UNNECESSARILY

APAR status

• Closed as fixed if next.

Error description

• The Big SQL Scheduler mantains a table-related cache that
  stores information about the table files. This information may
  take some time to be collected, especially for tables comprising
  many files. Once loaded it's accessed quickly to allow a faster
  compilation of hadoop related statements.
  In some cases, the information gets invalidated completely and
  is forced to be reloaded. There are however cases when this
  reload happens unnecessarily, as in the case of
  
  ALTER TABLE ... ADD PARTITION ...
  ALTER TABLE ... DROP PARTITION ...
  MSCK REPAIR TABLE ...
  
  The above commands are examples of commands that execute
  primarily in Hive, and do not need to have the table metadata
  available in the Scheduler cache. Reloading the cache
  unnecessarily in these cases can introduce a delay.
  In the case of tables with a few thousand files this can
  introduce a noticeable delay in the execution of the statement
  that may affect concurrent operations and generally slow down
  operations.
  

Local fix

• Execute the commands through Hive especially if several of them
  are needed on the same table then re-synchronize the table in
  bigsql.
  

Problem summary

• Please see problem description.
  

Problem conclusion

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

Fix information

Applicable component levels