Support Questions

Find answers, ask questions, and share your expertise
Check out our newest addition to the community, the Cloudera Data Analytics (CDA) group hub.

Install registered new HDF3.0 to existing cluster (repo hosted on RedHat satellite server)


Currently upgrading one of existing HDF 2.1.2 cluster.

Followed the instruction

  • finished all steps to upgrade Ambari related components (ambari server/ agent, infra, metrics, grafana etc and restart)
  • restart the cluster and passed the service check for all services

Now planning to upgrade HDF ( we have kafka installed in the cluster)

When registering the new HDF 3.0 version, because we are using satellite server to hosting the repo, I found I have to input full version number to save the version registration.

Our satellite repo is setup based on HDF 3.0 release note.

So, I think the version should be

By register the HDF- I proceed to install the new version on existing cluster.

Now, the steps of "Intall on... " have pending there for more than 1 hour, I manually checked the /usr/hdf folder on each of the box, there's no new version being copied on any of the cluster machines.

And I run yum info kafka and can tell that we have in the repo.(while the release note mentioned the kafka version is

Is that possible the version hosted on satellite server (set up according to release note) not matching the version I've given (I thought it's because all the version listed on release note is

So what I should do, because there's no cancel button to current installation and I don't know where to check the current step log.

Could anyone help by giving more detailed information on this upgrading step while using satellite server?




I de-registered the version and found nothing is done by the previous step.

After check, I found, take Kafka version as an example,

in release note, it's Kafka

in yum info (repo setup according to release note), it's kafka

in Ambari Web UI version registration, it's kafka

So I guess I might need a version definition file? Where to download this? thanks


Hi, Just to update the progress.

With the help from Hortonworks Support team, we got a workaround to this issue.

The reason causing this issue is because the initial installation didn't pull the stack version Satellite server which results in unsuccessful reading the current stack version during updating.

The workaround is to directly update the version information in Ambari configuration database. The workaround works for our environment and the stack is successfully upgraded to HDF3.0.

But still, could anyone give some clue on this issue. How Ambari get the repository version and OS information if it's a satellite server environment?

We want to find the root reason for this, but because Hortonworks do not have Satellite server environment to compare from, it's hard to solve this issue completely, we might have to perform the workaround again and again in different clusters and for different versions.

Take a Tour of the Community
Don't have an account?
Your experience may be limited. Sign in to explore more.