Parcel deployment on an existing cluster does not rely on SSH logins. Only installation of new agent hosts do, during which a form is presented to pass these credentials. The way we deploy parcels in CM is by having the CM server download it locally to its configured parcel repo disk, and then this file is sent to the agents via the HTTP service that CM runs (agents are instructed to download it from the CM server, which now acts as the repo). The entire procedure relies simply on connected agents, and involves no extra login requirements. It would be helpful if you can share a snippet of the message seen or explain your workflow more (especially if the failure you observe is consistent). ... View more

CM 5.5.3 is out now with a fix, and CDH 5.5.2 has been re-released. Read more at http://community.cloudera.com/t5/Community-News-Release/ANNOUNCE-Cloudera-Enterprise-5-5-2-and-Cloudera-Manager-5-5-3/m-p/37580#M107 ... View more

Setting the "oozie.action.launcher.mapreduce.job.ubertask.enable" property to "false" via the "oozie-site.xml" Service safety valve in CM can help serve as a workaround to this bug. The fix for this is otherwise included in CDH 5.5.2 and higher (in which the workaround is no longer required). ... View more

I prefer using the simpler bash syntax of using special escaped characters, if it helps: We know that ^M is the same as \r, which makes sense if you used Windows Notepad to write the commands but forgot to convert the file via dos2unix: ~> echo $'\x0d' | cat -v ^M ~> echo -n $'\x0d' | od -c 0000000 \r 0000002 (The \x0D or \x0d is the hex equivalent of \r, per http://www.asciitable.com/ (carriage return)) Therefore, you can use the $'' syntax to write a string that includes the escape: ~> hadoop fs -ls $'/a/b/c/d/20160206\r' Or, ~> hadoop fs -ls $'/a/b/c/d/20160206\x0d' This words well regardless of the terminal emulator you are using, cause we're escaping based on representation vs. by reliance on the emulator understanding the characters via input. ... View more

Its easy to get carried away by the words "namespace separation" and what it truly means in federation setup. Its not as simple as it sounds. If you have 3 NNs, you will have 3 NNs with "/" each. Its not 3 NNs with "/hbase" on all of them and files distributed automatically among them. ViewFS is a client-sided FS that can talk to the 3 NNs, and rewrite the path locally. That is, you can choose to mount / of NN1 as /hbase1, / of NN2 as /hbase2 and / of NN3 as /hbase3, but not bind all of them to a common root (/). When you then ask for viewfs:///hbase1/foo, it internally goes to NN1 looking for just /foo. NN1 does not know what "/hbase1" is, only the client (viewfs) does, which translates/brokers the request for you automatically. Simply said, no - such a setup with a single HBase cluster over 3x NNs is not possible. Even if you did provide some form of viewfs:// as the root directory in HBase config, it would fail cause of viewfs restrictions around moving paths between mount-point paths. ... View more

You are looking for HBase's request-throttling features: http://blog.cloudera.com/blog/2015/05/new-in-cdh-5-4-apache-hbase-request-throttling/ ... View more