问题
It seems that on spark 3.0.0 when I do a spark submit with kubernetes it require kerberos, I use the same spark submit that was working great in 2.4.5 I get this error:
Using Spark's default log4j profile: org/apache/spark/log4j-defaults.properties
20/07/04 08:17:51 INFO SparkKubernetesClientFactory: Auto-configuring K8S client using current context from users K8S config file
20/07/04 08:17:51 INFO KerberosConfDriverFeatureStep: You have not specified a krb5.conf file locally or via a ConfigMap. Make sure that you have the krb5.conf locally on the driver image.
Exception in thread "main" org.apache.hadoop.security.KerberosAuthException: failure to login: javax.security.auth.login.LoginException: java.lang.NullPointerException: invalid null input: name
To be more specific, I don't want to use kerberos
回答1:
Ok, found, you need to specify a spark_uid at image build step.
回答2:
either you handle this id ( 185) at os level, or you can comment it in the dockerfile but you will be root.... Eric
来源：https://stackoverflow.com/questions/62726949/how-to-launch-spark-3-0-0-kubernetes-workload-without-kerberos