About Adija1

Adija1 · ‎12-13-2016

@jss I probably didn't explain myself correct: The mysql error isn't gone - the mysql-connector-java.jar file is gone ! I copied it to the /usr/hdp/current/hive-server2/lib path but after restarting the Hive service - the file is gone! It is as if it is being deleted by the restart itself . Look at the screenshot of before and after (The shell in which i did "ll | grep mysql") - it's there, and after restarting - it's not there anymore. The file is gone. Vanished. Disappeared. Regarding the templeton.libjars here is the screenshot after the fix:

Adija1 · ‎12-13-2016

@jss 1. Corrected the path to /usr/hdp/${hdp.version}/zookeeper/zookeeper.jar,/usr/hdp/${hdp.version}/hive/lib/hive-common.jar Still same error in service check. Seems like a straight up bug cause it's searching for hive-common.jar/zookeeper.jar. 2. mysql-connector-java.jar was missing from /usr/hdp/current/hive-server2/lib I found one in /var/lib/ambari-server/resources/mysql-connector-java.jar and copied it to /usr/hdp/current/hive-server2/lib After restarting the Hive service - it was gone. I checked it twice. Here is the before & after using same command " ll | grep mysql" How can Hive work properly if it has problems connecting to the metastore ?

Adija1 · ‎12-13-2016

@jss I've compared to other working environments and the value is the same in all of them: /usr/hdp/${hdp.version}/zookeeper,/usr/hdp/${hdp.version}/hive/lib/hive-common.jar/zookeeper.jar

Adija1 · ‎12-13-2016

@Divakar Annapureddy Thank you for replying! The HIVE metastore is on the same host as the running HiveServer2. Nothing was changed in the hive mysql \ metastore- only Ambari was upgraded. That's the only change. Also - if there was a problem with connection between the HS2 to the metastore - Hive wouldn't work, but fortunately it works without a problem! I have NO problems with restarting Hive nor working with it. I can connect to any Hive DBs, query as usual. Permissions are working. Everything seems normal and green. My problem is that the restart indicator doesn't go away + restarting produces error although it succeeds with restarting + service check fails.

Adija1 · ‎12-13-2016

Hi @jss Thank you for your reply The "templeton.libjars" Property hasen't been changed so i'm kind of baffled why i need to "correct" it. I do agree that the path which the service check is looking seems corrupt: /usr/hdp/2.3.2.0-2950/hive/lib/hive-common.jar/zookeeper.jar hive-common.jar/zookeeper.jar (jar file \ jar file) It seems like a bug in the service check.... there can't be a .jar file under a .jar file... Or am i missing here something ?

Adija1 · ‎12-13-2016

Hello HDP 2.3.2 & Ambari 2.1.2.1 that was upgraded to Ambari 2.4.2.0 After upgrading Ambari to 2.4.2.0 when i start Hive service the operation output is red with an error - but Hive is green, up & running without any problems. Here is the error: But this is the actual result - the restart indicator still exists and Hive is green. I'm also able to connect to hive using beeline and work as usual. Service Check produces the following result The hiveserver2.log is free of errors and Hive is working properly ! I can live with this but i'm afraid that i will have problems with upgrading the HDP because one of the prerequisites is successful service check... Any ideas ?

Adija1 · ‎11-29-2016

Found the problem. Investigating the hiveserver2.log showed that rangerlogger failed to flush data to the db due to permission problem (Internal Exception: java.sql.SQLException: Access denied for user 'rangerlogger'@'<server>' (using password: YES) Error Code: 1045) After granting the permissions it flushed the data successfully. Now i just need to figure out what caused the loss of permissions and how the hell it is related to the server reboot.... Thanks for all your help !

Adija1 · ‎11-29-2016

@Edgar Daeds I've changed those settings to clear text and restarted hive + Ranger but problem remains unfortunately.

Adija1 · ‎11-29-2016

@Sagar Shimpi Can you be more specific ? The Ranger authentication is LDAP, meaning in order to login to it's UI you can only login using your Active Directory credentials. This works without any problems.

Adija1 · ‎11-29-2016

One more thing to add: we restarted the server after installing and configuring sssd on the server for group mapping.

Online	Offline
Last Visited	‎09-20-2018 02:06 PM

Member Since	‎02-04-2016 06:49 AM
Last Visited	‎09-20-2018 02:06 PM
Posts	132
Kudos received	52

Cloudera Community

Re: Connection failed to DataNode:50075 sometimes

Re: decommision datanode and keep other service

Re: Ambari-Agent high cpu & Datanode without heart...

Re: Permissions problem in Capacity Scheduler view...

Re: Ranger Audit stopped working after server rebo...

Re: HiveServer2 shows failure - but starts after u...

Re: HiveServer2 shows failure - but starts after u...

Re: HiveServer2 shows failure - but starts after u...

Re: HiveServer2 shows failure - but starts after u...

Re: HiveServer2 shows failure - but starts after u...

HiveServer2 shows failure - but starts after upgra...

Re: Ranger Audit stopped working after server rebo...

Re: Ranger Audit stopped working after server rebo...

Re: Ranger Audit stopped working after server rebo...

Re: Ranger Audit stopped working after server rebo...