Solr Cloud: Indexing in a Map reduce Job with Kerberos

[Bertrand Venzal <bv...@yahoo.com.INVALID>]

Hi all,

As a bit of background, we're trying to run a map-reduce job on a Hadoop cluster (CDH version
5.4.5) which involved writing from Solr during both the Map phase. To accomplish
this, we are using the Solrj library with version 4.10.3-cdh5.4.5. In the driver class which launch the MR Job, we have managed to get solrj to authenticate correctly with Kerberos,
creating a jaas file which points to a previously created keytab file. However, the distributed map-reduce job will not have the jaas file or keytab file on the worker nodes it will be distributed
to. We were wondering whether it is possible to set up Kerberos with Solrj in the driver step
of a map-reduce job, so that the individual mappers and reducers remain authenticated during
the running process?

Thanks and kind regards,Bertrand