Hi Artem, PHD-3.0 is equivalent to HDP-2.2.4, so a script referring to correct paths should work. What do you mean by "hdp memory script" on Yarn? Edit: phd-configuration-utils.py exists on PHD-3.0, here is a link, scroll down to "PHD Utility Script". Example python phd-configuration-utils.py -c 16 -m 64 -d 4 -k True where "-c" is the number of cores, "-m" is RAM on worker nodes, "-d" is number of disks, and final Boolean is HBase yes/no. ... View more

Hi @Rohit Sureka, I just tried (HDP-2.3.2 sandbox) and for me the timestamp works per documentation, with up to 9 decimal places (nanoseconds). Can you check your input data, and delimiters of your fields, are they as expected. Here is my test. My table: hive> create table ts(id int, t timestamp) row format delimited fields terminated by ',' stored as textfile location '/user/it1/hive/ts'; A few lines of my input file 11,2015-11-01 21:10:00 12,2015-11-01 21:10:00.1 15,2015-11-01 21:10:00.123 And a select/order by command hive> select * from ts order by t; OK 11 2015-11-01 21:10:00 12 2015-11-01 21:10:00.1 25 2015-11-01 21:10:00.1190011 37 2015-11-01 21:10:00.12 15 2015-11-01 21:10:00.123 31 2015-11-01 21:10:00.1234 17 2015-11-01 21:10:00.12345 19 2015-11-01 21:10:00.123456789 21 2015-11-01 21:10:00.490155 57 2015-11-01 21:10:00.60015589 Time taken: 2.34 seconds, Fetched: 10 row(s) ... View more

Hi @Rajendra Vechalapu, you need to login as root to your Sandbox or a machine where you want to install Spark and run the wget command there. It's also assuming that your OS is CentOS or RHEL-6. I just tried and it works (the URL is still valid). After that run "yum install" and follow all other steps on that page before trying Spark from the command line (test the command line first) or from Zeppelin. ... View more