Ok, I'll look into it, please stand by.
In the meanwhile you can use a specific version like xenial-cdh5.15.1 instead of xenial-cdh5 in your sources.list to workaround this issue.
I think what is happening is that CM Server is running 5.16.0. When adding a cluster, the hosts need to run the CM agent that matches that version. Since it does not exist, the install is failing.
Only thing to do, since there is no 5.16.0 release anymore, is to wait for CM5.16.1 release. Upgrade CM to that release, then your agents will look for 5.16.1 and will find them.
I would WAIT until 5.16.1 is announced even if you happen to see it on archive.
Sorry, this is the only way I can think of to get your install finished.
dpkg --get-selections | grep cloudera
dpkg -l | grep cloudera
Those commands on Ubuntu should show you the version of CM installed I don't know of any other command to show if the service is running or stopped off hand.
It is important that we understand what steps you are doing to upgrade things in your cluster. Can you provide the steps you are doing to update things?
If you are using apt-get, it sounds like further debugging needs to be done there to figure out why it thinks 5.16 exists. 5.16 references have been removed from Cloudea server
Perhaps strace could be used to check what files are being accessed before attempting to download 5.16.
If you have 5.16 bits but have not restarted Cloudera Manager so it is still at 5.15.1, you can stop CM, remove the 5.16 package and then install 5.15.1. Cloudera Manager does not do anything to the CM database unless CM is started with the new binaries, so it should be safe to "roll back" in this case. Of course, make sure you have a backup of your database before trying it (just to be safe).
The result of the commands:
dpkg -l | grep cloudera ii cloudera-manager-agent 5.15.1-1.cm5151.p0.3~xenial-cm5 amd64 The Cloudera Manager Agent ii cloudera-manager-daemons 5.15.1-1.cm5151.p0.3~xenial-cm5 all Provides daemons for monitoring Hadoop and related tools. ii cloudera-manager-repository 5.0 all Cloudera Manager ii cloudera-manager-server 5.15.1-1.cm5151.p0.3~xenial-cm5 all The Cloudera Manager Server rc cloudera-manager-server-db-2 5.14.2-1.cm5142.p0.8~xenial-cm5 all This package configures an "embedded" PostgreSQL server,running as user cloudera-scm on port 7432.
As how did this happen, I was just simply trying to update/upgrade the Ubuntu server which is hosted CM. I wasn't aware of a new version so I wasn't trying to upgrade the cluster at that moment (it was purely for security reason to have up to date packages).
Since the upgrade never happened due to the errors from the beginning, I am wondering how is it possible to have any 5.16.x bits in my system?
Please note: The first release of Cloudera Enterprise 5.16 is numbered 5.16.1.