Setting your entire HDFS to 777 perms is not recommended. You just needed to make sure there was a 777 /tmp directory. CM normally creates this for you as part of HDFS setup, and there's a command in HDFS to create the /tmp dir if you need to run it manually.   It's not really convenient to undo this, so if you're fine with having effectively no hdfs permissions, then you can keep things as is. Otherwise you can manually re-restrict permissions to each folder appropriately. Anybody else hitting this issue should just create /tmp instead of the recursive chmod. ... View more

Glad you got it working! ... View more

Assuming there's nothing valuable in your NameNode, try deleting your NameNode data directories and retrying your first run. There should be a retry button on the page that says First Run. You can find the namenode data directories on the configuration page in the wizard, or by clicking on HDFS and viewing the configuration. It may also help to see the stdout and stderr log of the NameNode format command, which you can find by clicking on HDFS, then commands. ... View more

Since you re-created a cluster on the same hosts, HDFS was already formatted from your previous run. Your overall workflow should have succeeded and you can ignore this error. If all your services started, then you're good to go. 544 is the command ID in the CM database, not an error code. It's basically for debugging. ... View more

Your username / password don't match. Easiest thing to do is delete and re-add the Hive service, which will create a new database account for you. I'm not sure why that didn't work in the first place. If that doesn't work, then change the database name to "hive2", the username to "hive2", run the command to create the Hive database, then the command to create Hive tables, and you should have a brand new schema that the HMS can start on. ... View more

If you were just adding ZooKeeper, then it should not prompt you for any database test since ZooKeeper doesn't have a database. For steps 2-4, did you get all the way through the wizard, to the point where the services were started and the wizard returned you to the home page? What database connection did it prompt you to test when adding ZooKeeper? There should be a name for the connection, like Hive or Activity Monitor. ... View more

Hi,   It's not clear to me exactly what you were doing when you hit an error. Here's what I think you did, with questions where I'm not sure:   1) Installed CM and CDH binaries on all hosts - Did you use packages or parcels to install CDH? I'm guessing parcels, otherwise your later CDH install by packages wouldn't have done anything. 2) Installed HDFS and YARN on the cluster (at the same time) 3) Did NOT complete the wizard. HDFS, YARN, and the Mangement service were NOT started 4) Hit the back button and tried to add ZooKeeper as well, then hit the problem in AddServiceWizardController2. 5) Manually installed cdh binaries. You should not have done this, since it will conflict with whatever you did in step 1 for CDH binaries. If my assumptions are correct, you now have CDH installed as both parcels and packages, which will cause problems. You'll need to yum erase all of the packages you installed.   Keep in mind that installing binaries is completely separate from configuring services.   Assuming I'm correct, then probably step 3 and 4 left your cluster in a partially completed state. You can click on the Cloudera Manager logo to get to the home page, delete your cluster from the dropdown menu to the right of the cluster name, then click the dropdown menu near the word Status and select the Add Cluster option to start over again. If you have both parcels and packages installed on that cluster, you'll need to manualy uninstall your packages.   To configure a service to run on an existing cluster, click the dropdown menu to the right of the cluster name and select Add Service. This will let you configure things like ZooKeeper or Hive to run on an existing cluster.   Thanks, Darren ... View more