In that case, my suggestion is to check for HS2 logs to try to find any queries that takes time to compile. Because HS2 currently can only compile one query at a time, if you have a query that takes a while to compile, it will lock out HS2 and other requests will have to wait. I am not sure if it is the cause, but worth to check the HS2 log. Look for the following example: Completed compiling command(queryId=hive_20170227192424_ee0c5216-bef5-456a-934e-212aabe55b75); Time taken: 0.191 seconds and check if any Time taken values are too big around the time when issue happens. ... View more

Hi,   I can see the images now, but not sure why they did not show up before. I can see that you are trying to run a query that contains:   (rel.i_prtitn,rel.i_mod_yr) in (select * from my_prtitn)   I would like you know that this is currently not supported, please refer to Hive's Apache documentation below: https://cwiki.apache.org/confluence/display/Hive/LanguageManual+SubQueries   It mentions the following:   IN/NOT IN subqueries may only select a single column If you want to acheive the same goal, you will need to use JOIN with subQuery on those two columns.   Hope this helps.   Regards Eric     ... View more

Yes, HMS and HS2 have separate log4j.properties files. However, I find it odd that one cluster has the threadname in the log records while the other does not (both CDH5.4.7). The defaults should be exactly the same on both the clusters.   Perhaps, you could confirm that you have not overridden the configuration in a safety valve.   To add a threadname to the log records for the HMS, Select Hive Service --> Instances --> Hive Metastore Server --> Configuration then type in log4j and look for "Hive Metastore Server Logging Advanced Configuration Snippet (Safety Valve)" This is where I would set it to something like this .. log4j.appender.RFA.layout.ConversionPattern=%d{ISO8601} %-5p [%t]: %c{2} (%F:%M(%L)) - %m%n or just %d{ISO8601} %p [%t]: %c: %m%n   the %t is what adds the thread name to the log records.   Hope this helps. Thanks ... View more

If this is from HUE, hue can execute multiple command in a sequence until it reaches the first query that returns results (like a select query). For example, the following should be executed entirely in a single go.   drop table if exists foo; create table if not exists foo (code string, description string, salary int); insert into foo select code, description, salary from sample s where s.salary > 50000 and s.salary < 100000; select * from foo where salary < 75000;   The following will stop after the select query, so the drop table will not be executed. drop table if exists foo; create table if not exists foo (code string, description string, salary int); insert into foo select code, description, salary from sample_07 s where s.salary > 50000 and s.salary < 100000; select * from foo where salary < 75000; drop table foo;   But if you use beeline to execute a file containing multiple select queries, this should work without pausing. ... View more