T
Tim Kirwan
Guest
Additional information regarding the business case. Some customers of an OEM have very large hadoop clusters. They are using the OEM's JDBC Connectivity capabilities for reading from hive. Kerberos is enabled in their environment and with the current support for the kerberos that is provided by the driver, the keytab or the cache file needs to be created on each node. With a large cluster size of ~120 servers, it is virtually impossible for the customer to create the keytab and cache files and the required configuration files on all the nodes in the cluster. In order to address the same, the OEM has provided a capability to localize required files on all the nodes. This works fine in case of the ODBC. But, in the case of the JDBC, since it reads the configuration file and does not support the Kerberos environment variables, the OEM is running into challenges in localizing the keytab and cache files specified in the configuration file. Any alternative solution on that the OEM could do would need good amount of changes and hence it would be very much beneficial for us if the driver supports this particular functionality.
Continue reading...
Continue reading...