question Re: How to determine who ran query if Hive Impersonations = False? in Archives of Support Questions (Read Only)

How to determine who ran query if Hive Impersonations = False?

sunile_manjee — Wed, 20 Apr 2016 09:23:03 GMT

I have hive impersonations set to false. Is there a simple way of determining who ran the job/query?

Re: How to determine who ran query if Hive Impersonations = False?

aromero — Thu, 21 Apr 2016 05:46:36 GMT

Hi Sunile,

You can use a Hive hook to determine who submitted the query, even with doAs=false. As an example you can even use that to direct the query to an specific YARN queue. See the following for more info:

https://github.com/beto983/Hive-Utils/blob/master/README.md

Re: How to determine who ran query if Hive Impersonations = False?

sunile_manjee — Sat, 23 Apr 2016 02:42:29 GMT

Thanks to @Guilherme Braccialli I found it can be done via tez api calls.

some examples thanks to @Guilherme Braccialli

http://myserver.com:8188/ws/v1/timeline/HIVE_QUERY_ID/hive_20160330062211_6917b0b2-a34f-4735-b332-57083a18199f

http://myserver.com:8188/ws/v1/timeline/HIVE_QUERY_ID/hive_20160330064546_192ee370-d548-44fc-8073-903eaf6bf02

Re: How to determine who ran query if Hive Impersonations = False?

gbraccialli3 — Sun, 24 Apr 2016 08:03:17 GMT

@Alberto Romero

Notice that if you are re-using initialized sessions (hive.server2.tez.initialize.default.sessions) this solution will not work, because the properties are created only when the AM starts. When you use default sessions the AM will be created before your hook can define a custom property.

The hook solution work for non-default tez sessions and also for map-reduce.

Find here some rest calls to RM API that works for all Tez scenarios (but not for map-reduce engine).

Re: How to determine who ran query if Hive Impersonations = False?

aromero — Sun, 24 Apr 2016 14:15:41 GMT

@Guilherme Braccialli Indeed, the hook is for not reusing initialised sessions.

Thanks very much for sharing the REST calls. Look quite useful to have handy.