Details
-
Bug
-
Status: Resolved
-
Minor
-
Resolution: Information Provided
-
1.4.6
-
None
Description
I am importing data from SQL Server to HDFS. Seems there is a bug when autnetication mode is LDAP and format is parquet. Please let me know if this is a bug or I am doing incorrectly. I am using Cloudera distribution.
Below different scenarios I have tested
Command Used are.
1) SQL Server (Native Connection) -> HDFS Text Format (WORKS FINE)
sqoop import \
--connect "jdbc:sqlserver://<<ServerName>>;databasename=<<DB_Name>>"
--username xxxx \
--password xxxx \
--table 'table1' \
--split-by col1 \
--target-dir /user/user1/01/
2) SQL Server (Native Connection) -> HDFS Parquet Format (WORKS FINE)
sqoop import \
--connect "jdbc:sqlserver://<<ServerName>>;databasename=<<DB_Name>>"
--username xxxx \
--password xxxx \
--table 'table1' \
--split-by col1 \
--target-dir /user/user1/01/ \
--as-parquetfile
3) SQL Server (LDAP Connection) -> HDFS Text format (WORKS FINE)
sqoop import \
--connect "jdbc:jtds:sqlserver://<<Server2>>;useNTLMv2=true;domain=XXXX;databaseName=<<DB_name>>" \
--connection-manager org.apache.sqoop.manager.SQLServerManager --driver net.sourceforge.jtds.jdbc.Driver \
--username xxxx \
--password xxxx \
--table table2 \
--target-dir /user/user1/02/ \
– --schema=s1
4) SQL Server (LDAP Connection) -> HDFS Parquet format (Executes successfully, but the exported files are not parquet, they are in TEXT format)
sqoop import \
--connect "jdbc:jtds:sqlserver://<<Server2>>;useNTLMv2=true;domain=XXXX;databaseName=<<DB_name>>" \
--connection-manager org.apache.sqoop.manager.SQLServerManager --driver net.sourceforge.jtds.jdbc.Driver \
--username xxxx \
--password xxxx \
--table table2 \
--target-dir /user/user1/02/ \
– --schema=s1 \
--as-parquetfile