You may install a Spark parcel using the CM instance inside the current VM:
Our next version of the VM, for Cloudera Enterprise 5, will contain Spark natively.
No, since there isn't something to upgrade 😉
You're creating a new vm from scratch. You do not need to take the latest available quickstart vm as a starting point.
sorry, but i really don't understand the instruction of 'build your own CDH5' . it seems it has assumptions the reader has some experience with centos. i tried to ask google where i didn't have the answer in the instruction but i feel hopeless..
In addition, I tried to install spark parcel over the quick vm - but when i asked to update the envoirment params in the spark_env.sh - > i set the HADOOP_HOME as i asked to but still had problems with running the master (remember - i use only 1 node).
seems like you're thinking too complicated 😉
You are not forced to use CentOS. If you are familiar with Ubuntu, just create a new vm with plain Ubuntu installation. Afterwards install ClouderaManager5 (like described here) and choose installation Path A.
To put it in a nutshell:
1.) create a new vm with base OS installed
2.) ensure prerequisites are fulfilled before installing ClouderaManager5 (see link above)
3.) install CDH5 including Spark by using ClouderaManager on your singlenode "cluster" (==the recently created vm)
Or, btw., since CDH5 GA is available since yesterday possibly there will be an upgraded quickstart vm released soon ?!?
ubuntu -> trying to install the cloudera manager and got this message suddenly.
i checked the log :
then i tried to install postgresql but still i had the same issue.. 😞
ok i step forward in the installation.. now my problem is what to set in the host ip.. i tried 127.0.0.1 but it can't add it cause it can't connect via ssh...