What kind of role/access control can a Cloudera Admin make/give, so that an end user, who is doing ad-hoc querying has ability to access the Impala Web User Interface described here   https://www.cloudera.com/documentation/enterprise/5-3-x/topics/impala_webui.html   Our cordinator nodes throw a user id and password dialog and my user id has no permission, so trying to figure out what to ask our admins to open up and is it even possible. Their concern was the Cancel button that is shown in this webui against running queries and anyone having the ability to cancel queries being run on the coordinator node.   I have no interest in that, have never done that, and really interested in the below pages of that webUI   http:// impala-server-hostname :25000/memz http://impala-server-hostname:25000/queries http://impala-server-hostname:25000/sessions   Queries page I believe has the graphical explain plan, which is the most important one that I need to see and like over the textual EXPLAIN plan that I can see by running a EXPLAIN command in my query tool. Memz page is also usefull to see if Cluster memory utilization is high leading to ETIMEDOUT errors when running queries.   So is there a way for an end user to access it and for queries page, if not all running queries, atleast have ability to see queries he/she executed.   Ideal would be CM interface, so that I don't have to go to each coordinator and check if my query is running there...as described here....but not sure if CM is restricted to Administrators " To get a convenient picture of the health of all Impala nodes in a cluster, use the Cloudera Manager interface, which collects the low-level operational information from all Impala nodes, and presents a unified view of the entire cluster. "   From looking at roles available in CM, what would give access to Impala Web UI with cancel button disabled for everyone.   https://www.cloudera.com/documentation/enterprise/5-3-x/topics/cm_sg_user_roles.html     ... View more

We have a strange issue going on in our 5 node cluser (256 gb RAM, with 130 gb assigned to Imapala in each node). We have been unable to run a particular query that is joining quite a few tables. It also creates 5 temporary tables that is ultimately joins to create a final query output that is supposed to be passed back to Informatica BDM.   But as soon as the query is sent from Informatica, it finishes running and returns no output. This behaviour is not observed in our pre-prod environment, but that again does not have that much data. What our Cloudera admin can see is that the query is coming from Informatica, but he sees a LIMIT 0 added to the end, which we do not have in Informatica mapping. That then likely causes no output to be returned.  He sees similar thing for a different query from a different mapping, but in that case, as second query is seen without LIMIT 0 and we do get data returned for that Informatica mapping.   So we are are just wondering if we have a cluster configuration issue or does Impala issue a limit 0 query first to get the EXPLAIN plan and then issues the real query, but in this case, it is failing to do so for some reason.   When I ran the EXPLAIN command around the complex query from our query tool (TOAD Data point connected via Impala ODBC Driver), I get this   Per-Host Resource Reservation: Memory=3.85GB Per-Host Resource Estimates: Memory=9.83GB   I do not have expertise to understand rest of the 1000 line EXPLAIN commands output to understand what is the total Impala RAM that this query is demanding.   Any pointers from the community experts is much appreciated ... View more

Just installed Quickstart 5.13 as a Docker container running in Docker for Windows on Windows 10. Have started going through the Quickstart Tutorial. Just wondering if there is a way to persist the Tables and Data created in these Tutorials, so that I don't have to redoo for instance the Sqoop script in Tutorial one to create the tables.   I have created a Shared Volume on my Windows Drive to use and be available for use by Quickstart Container in the /src folder.   docker run --name=my-cloudera --hostname=quickstart.cloudera --privileged=true -t -i -v /Users/UserName/Documents/DockerShare:/src --publish-all=true -p 8888:8888 -p 80:80 -p 7180:7180  cloudera-5-13 /usr/bin/docker-quickstart   Not sure beyond this, how to have data for instance in HDFS at user/hive/warehouse/ to get persisted between container re-start and if that is even possible.   I could have gone the Oracle Virtualbox Quickstart VM route which I have done 2-3 years back, but just learnt about Docker and find it to take far less resources and start everything up so fast. ... View more