I think you’re missing a step. You need to uncompress the tar.gz
Quoting from the Prism Web Console Guide:

1. Perform these steps for NCC versions that include an installer file inside a compressed tar file (ncc_installer_filename.tar.gz).
   1. Extract the installation package.

      nutanix@cvm$ tar xvmf ncc_installer_filename.tar.gz --recursive-unlink 

      Replace ncc_installer_filename.tar.gz with the name of the compressed installation tar file.

      The --recursive-unlink option is needed to ensure old installs are completely removed.

   2. Run the install script. Provide the installation tar file name if it has been moved or renamed.

      nutanix@cvm$ ./ncc/bin/install.sh l-f install_file.tar]

      The installation script copies the install_file.tar tar file to each node and install NCC on each node in the cluster.

The numbering didn’t cooperate bringing over the quoted text. This is step 4 in the doc. 

Generally speaking I’d recommend to stick with the installer.sh where possible. You should usually be able to wget this from the CVM or download it through your browser by just following the URL format and specifying the version you want, like is described in this KB

Hi Jeremy

Thanks for your reply.

Of course I uncompressed the tar.gz file before run the script. Sorry I didn’t provide the screenshot because of the uncompress process is long.

The process is useful for us to install NCC on each node in the cluster. What is potential causes to show “Installation of ncc failed on node xx.xx.xx.xx”? Is node data transmission issue? How to troubleshoot the issue?

By the way, I have tried to ping each node successfully from cluster.

Hello @Danny.H 

Just to validate. 
One of the steps mentioned in the guide is to rename the tar.gz file. 
Did you perform that action?  

 

1. Replace ncc_installer_filename.tar.gz with the name of the compressed installation tar file.(This task specifically)

    

2. Run the install script. Provide the installation tar file name if it has been moved or renamed.

   nutanix@cvm$ ./ncc/bin/install.sh n-f install_file.tar(not ncc_installer_filename.tar.gz)]

Hi HITESH0801

Thanks for your reply.

I tried to unzip the ncc_installer_filename.tar.gz to ncc_installer_filename.tar file under Windows, and then upload it to the cluster. 

But the result is the same as below.

I also tried to rename the ncc_installer_filename.tar.gz to ncc_installer_filename.tar directly. The result is the same too. 

By the way, I can type below command to install script on local CVM successfully. 

./ncc/bin/install.sh -l

Hello @Danny.H 

Any particular reason you’re going ahead with the tar method?
Single installer file can also install NCC on each node or even single(in some scenarios ) and is recommended if using 1-click is not an option. 

 

@Danny.H ,

Sorry I missed the mark earlier. You’re not running the command wrong. This is a (new?) bug. 

The local install option with the .tar.gz works but the global install from the tar.gz file method isn’t completing. 

If you want to install on just one node, use the .tar.gz method with the -l option. If you want to install/update NCC on all nodes without running the install once-per-node, for now there’s the Prism 1-click method and the ./ncc_installer_filename.sh method available and working.

Hi JeremyJ 

I will use a single installer file or use the .tar.gz method with the -l option to install each node.

But I still recommend to fix the global install feature. It is useful for the other user who have several nodes to be installed.

I certainly agree. A ticket has been raised with the NCC team to address this. 

If your question here has been sufficiently answered I would appreciate if you tag the best answer.