Quanlong Huang created IMPALA-14236:
---------------------------------------
Summary: Measure the heap size requires to start catalogd on a
given scale
Key: IMPALA-14236
URL: https://issues.apache.org/jira/browse/IMPALA-14236
Project: IMPALA
Issue Type: Task
Reporter: Quanlong Huang
Assignee: Quanlong Huang
In the startup process, catalogd loads some basic metadata from Hive Metastore:
* All database objects
* All Java UDFs
* All table names
Note that all native functions are also loaded but they are contained in the db
parameters using keys with prefix "impala_registered_function_".
We'd like to know for a given scale of the warehouse (i.e. number of dbs,
tables, native/java udfs), how large of the JVM heap size is required for
catalogd to start. This should account for the intermediate objects that only
live in the middle of the operation, e.g. some objects returned in HMS RPC
responses.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
