Impala workload management table maintenance
Understand the maintenance requirements for the sys.impala_query_live and sys.impala_query_log tables.
For efficient query performance, different maintenance needs apply to the
sys.impala_query_live and sys.impala_query_log tables.Sys.impala_query_live:- No maintenance is required because it resides entirely in memory.
Sys.impala_query_log:- As an Iceberg table, it requires periodic maintenance, such as:
- Computing statistics.
- Optimizing the table structure.
- Performing snapshot expiration or cleanup.
- As an Iceberg table, it requires periodic maintenance, such as:
Since Impala workloads are unique, no automatic maintenance is performed on the
sys.impala_query_log table. You should schedule maintenance tasks according to
your workload needs.
To optimize the Impala query log, run the query OPTIMIZE TABLE sys.impala_query_log
(FILE_SIZE_THRESHOLD_MB=128). Cloudera
recommends testing this query in the development or test environments to evaluate its impact
on your workload. For best results, run the query during low cluster activity times.
