The host inspector pretty clearly indicated exactly which components were detected as CDH 4. For each of these, make sure the old versions were correctly uninstalled. Whatever steps you are following to uninstall is not working, possibly because you are using the wrong steps, and possibly because your package manager is in a weird state not anticipated by the uninstall instructions. Thanks, Darren ... View more

If you want /etc/hive/conf to be populated, then place Hive Gateways on those hosts. It's common to place it on all hosts so "hive" works on all hosts from command line. ... View more

Hi Sky, Host inspector showed that you had inconsistent versions of CDH installed. Log in to each machine and consult the package manager (such as yum for rhel-based systems) to ensure these were all correctly removed, or are the correct version. If you are using Parcels to install CDH (which is recommended), then you should have no CDH packages installed on your operating system. Thanks, Darren ... View more

This usually happens when you have manually installed CDH, but picked the wrong CDH version in the CM UI. Exit the wizard (click the Cloudera Manager logo), click the Hosts tab, and run the Host Inspector. It should say what versions of CDH are installed. If those are incorrect, fix them. Then run the Add Cluster wizard again (if you see an empty cluster on the home page, you may need to delete that cluster first), and be sure to pick the correct major CDH version. Thanks, Darren ... View more

Hi Mike,   You don't need to get them to a not-managed state. Instead, click the check box next to the hosts you want to use (or click the check box on the top to select all). Then the wizard will let you continue and configure those hosts with HDFS, YARN, etc.   Thanks, Darren ... View more

I'm sure you realized this by now, but deleting the entire log folder is definitely a bad idea. You should clean out the files that are troublesome, not remove whole directories. It's also good to configure data retention policies wherever you can. All Cloudera Management roles (like host monitor, activity monitor, and service monitor) have configuration options to control how much data they keep. I think when you are asking to increase the postgresql volume, you are talking about increasing the disk space available to /var/lib/cloudera-scm-server-db/data. The best thing to do is migrate to an externally managed database. If that's too difficult, then you can either alter the partitions on your machine to allocation more space to the relevant partition, or you can create a symlink from /var/lib/cloudera-scm-server-db/data to a place that has more available space. You should of course move the existing data to that new location before creating the symlink. Note that if you do the symlink strategy, it is likely that you will need to re-create it when you upgrade cloudera manager in the future, as the symlink confuses the RPM code. You won't lose your data on cm upgrade, just the symlink. ... View more