About Darren

Darren · ‎04-01-2015

Your default replication factor is probably 3. Since you only have one node, that's impossible to satisfy. On single-node clusters, you should reduce the replication factor to 1 in HDFS configuration in Cloudera Manager. Note that this only affects newly created files. Your existing files will still try to replicate to 3 hosts unless you change them explicitly (you should be able to google how to change a file's replication factor). Thanks, Darren

Darren · ‎03-16-2015

If you click on the Host Monitor and look at the role logs, stderr logs, and stdout logs, do they say why it failed to start?

Darren · ‎03-16-2015

From the home page, there should be a button in the upper right of your screen to add the management service. It only appears if no management service exists.

Darren · ‎03-12-2015

Look in the Resource Management category on the left of the configuration page. Each role group (there's usually one for each role type) has this subcategory. You can also search for "memory" or "heap" using the search bar on the left. Editing /etc/spark/conf won't help you since the daemon roles don't read that file. CM daemons don't read from /etc, but instead have their own independent configuration directories. http://blog.cloudera.com/blog/2013/07/how-does-cloudera-manager-work/ Thanks, Darren

Darren · ‎03-08-2015

Hi Balta, If you want your edge node to talk to YARN, add a Yarn Gateway role to that host. If you want it to talk to Hive, add a Hive Gateway. Etc. When you deploy client configuration, it'll go to all roles of that service, including gateways. Thanks, Darren

Darren · ‎03-05-2015

When using Spark on YARN, there's no need for the Master or Worker roles.

Darren · ‎02-26-2015

Hi Gregory I presume that the urls added back are among: http://archive.cloudera.com/accumulo/parcels/1.4/ http://archive.cloudera.com/accumulo-c5/parcels/latest/ http://archive.cloudera.com/sqoop-connectors/parcels/latest/ http://archive.cloudera.com/spark/parcels/latest/ That happens because certain services implemented with the CSD extension have a parcel repo associated with them, and they are always re-added on server start up. We will work toward fixing that in a future release. Sorry for the annoyance.

Darren · ‎02-24-2015

Files in /etc/ are not used by processes managed by Cloudera Manager. See here for more info: http://blog.cloudera.com/blog/2013/07/how-does-cloudera-manager-work/ Thanks, Darren

Darren · ‎02-20-2015

The activations are probably not completing successfully, making it stuck. Check agent logs for any clue to why activation's failing. Restarting all agents may help with resetting internal state and re-trying activation. If all of that fails, you can force de-activate the parcel using the API. http://cloudera.github.io/cm_api/ Thanks, Darren

Darren · ‎02-18-2015

I'll make sure we get parquet, hbase, zookeeper, and solr on that doc page. Thanks!

Online	Offline
Last Visited	‎05-21-2019 01:27 PM

Member Since	‎07-30-2013 10:59 AM
Last Visited	‎05-21-2019 01:27 PM
Posts	509
Kudos received	112

Cloudera Community

Re: Install using CM of Datanodes with different n...

Re: Cloudera API doesn't return any roles for a ho...

Re: Stopping selected roles from Service action me...

Re: CSD Role with Multiple SSLServers

Re: CSD - SSLServer Paramaters

Re: Percentage under replicated blocks: 100.00%

Re: Unable to issue query: could not connect to th...

Re: Unable to issue query: could not connect to th...

Re: how to increase spark's worker memory from CM

Re: Custom configuration in core-site.xml

Re: required help to install spark on CDH5

Re: "Remote Parcel Repository URLs" is always init...

Re: zookeeper datadir name configuration

Re: How do I fix KEYTRUSTEE stuck in "Activating" ...

Re: Can the CDH5.3 virtual machine be upgraded?