If its a new cluster, and its the yarn job, please make sure that you have mr-framework uploaded to /user/yarn/mapreduce/mr-framewor in hdfs. All yarn job will fail with below error if that path missing: Error: java.io.IOException: Can't get Master Kerberos principal for use as renewer $ hdfs dfs -put 3.1.1.7.1.7.2032-1-mr-framework.tar.gz /user/yarn/mapreduce/mr-framework ... View more

jus restart impala role and fix this,bucause there have manny tcp conntion ... View more

@araujo  do you have any suggestion for this case ?   ... View more

@iamfromsky as this is an older post, you would have a better chance of receiving a resolution by starting a new thread. This will also be an opportunity to provide details specific to your environment that could aid others in assisting you with a more accurate answer to your question. You can link this thread as a reference in your new post.  ... View more

oh, this is a long time ago issue, the root cause is because new machines charset is not utf-8, just keep all the machines chaset is utf-8 , then its ok. ... View more

i found there are some people using SSSD to resolve this issue , install SSSD on every machine, yarn service will find user in local os , if os doesn't have this user, it will find in openldap.  i have tested this solution, it works fine.   but i still don't want to install SSSD on every machine, so my question is still why hdfs, hive, sentry can work fine with openldap, but yarn is not . what should i do ? ... View more

in my pervious experience, i haven't ever set the port range, he default port is 32768 ---65536. so my only question is why 1000~ port can't be connected ?  could you give me some information? ... View more

Hello   A lot details are provided here to your questions https://docs.cloudera.com/cdp-private-cloud-base/7.1.6/concepts/topics/cm-sdx-vpc.html   Just to highlight a few points 1. not all components are available for VPC, limited to the following (no Flink) Only the following services can be installed on Compute clusters: Hive Execution Service (This service supplies the HiveServer2 role only.) Hue Kafka Spark 2 Oozie (only when Hue is available, and is a requirement for Hue) YARN HDFS (required) 2. data will be transferred to VPC during runtime, so check the network, performance requirements etc 3. only certain versions of the platform supports VPC, limited to the following: CDH 5.15 CDH 5.16 CDH 6.0 CDH 6.1 CDH 6.2 CDH 6.3 Cloudera Runtime 7.0.3 Cloudera Runtime 7.1.1 and higher 4. it is good you need to add temp compute resources to your cluster for some ad-hoc workloads, but it has the trade-off of the performance (refer to point 2) 5. VPC can be fully managed using Cloudera Manager, so it is a fairly easy process to implement ... View more