@vsrikanth9    Your krb5.conf entry is wrong  please change it to match the below   [domain_realm] .hadoopsecurity.com = HADOOPSECURITY.COM hadoopsecurity.com = HADOOPSECURITY.COM   The restart the kdc and kadmin # systemctl start krb5kdc.service # systemctl start kadmin.service That should resolve your problem Happy hadooping        ... View more

@irfangk1  There is something wrong I don't see the database entry in your ambari.server.properties  how will it bind ? Something like below server.jdbc.url=jdbc:postgresql://<HOSTNAME>:<PORT>/ambari?ssl=true Can you validate, you are using the embedded postgres right?   ... View more

@Gcima009    When you generate templates in NiFi, they are stripped of all encrypted values. When importing those templates into another NiFi cluster, Check your node that is not starting? has any values in the below parameters? You will have to populate all the processor and controller tasks passwords manually. Backing up flow.xml.gz or flow.tar file will capture the entire flow exactly as it is, encrypted sensitive passwords and all. NiFi will not start if it cannot decrypt these encrypted sensitive properties contained in the flow.xml. When sensitive properties e.g passwords are added they are encrypted using these settings from your nifi.properties file: # security properties # nifi.sensitive.props.key= nifi.sensitive.props.key.protected= nifi.sensitive.props.algorithm=PBEWITHMD5AND256BITAES-CBC-OPENSSL nifi.sensitive.props.provider=BC nifi.sensitive.props.additional.keys= In order to drop your entire flow.xml.gz or flow.tar onto another clean NiFi, these values must all match exactly.   Ref: http://www.contemplatingdata.com/2017/08/28/apache-nifi-sensitive-properties-need-know/ ... View more

@irfangk1    Can you share your ambari.server. properties and your /etc/hosts ... View more

@ThanhP  Good everything is perfect for you now 🙂 You ONLY execute sudo -u hdfs hdfs namenode -format  as a last resort because it's dangerous and not recommended to run that on production cluster as that [re-initializes] formats your Namenode hence deleting all your metadata stored on the NameNode.   Having said that the answer you accepted can't help a member who encounters the same issue "HDFS NameNode won't leave safemode" maybe you should un-accept it and accept your own answer as it's the more realistic answer . Happy hadooping    ... View more

@ThanhP  I can see you are accessing the web shell UI on 192.168.1.37 yet the sandbox is pointing to 172.18.0.2:8020. Now if you type in your browser http://192.168.1.37:1080/splash2.html you should land on the splash screen Choose the Ambari UI link , this should give you the normal Ambari user UI if you reset your password before as I had stated please use that admin/[password] combination and you will realize that none of the components are started HDFS, YARN, MR, HIVE, HBASE etc Note: some components are in maintenance mode unless you want any of them beware of the dependencies e.g ATLAS must have HBASE running etc.. that could explain why you got the error "Failed on connection error java.net.Connection" At a certain point during your startup, "Timeline Service V1.5" will error out with the message to do with "safemode on" get to the shell as hdfs $ hdfs dfsadmin -safemode get This should indicate it's ON so get it out of safemode $ hdfs dfsadmin -safemode leave And again use the option Ambari start all it will pick up from the last point of failure, after the components startup you can simply access them eg the name node UI will automatically point to http://sandbox-hdp.hortonworks.com:50070/dfshealth.html#tab-overview Please do that an revert ... View more

@ThanhP    What @Herman  was trying to explain is what happens in the background and you cannot influence or manually replicate the process.  The screenshots I attached are the best you can get to resolve the issue because the screenshots are from exactly the same sandbox you are using.   So follow those steps and revert ... View more