NetVault 13.3.2 - Administration Guide

Copy the packages to a shared location: Copy the client and plug-in packages to a shared location. Only CIFS shares, including Linux Samba Shares, are currently supported as package stores. The path must be accessible to the NetVault Server and all target machines where you want to install the packages.

Ensure that you use the original names of the installation packages. Renamed packages cannot be selected for push installations.

•

Configure a package store in NetVault: After copying the installation packages, configure the shared location details in NetVault. For more information, see Configuring a package store.

•

Configure the preferred network address setting on multihomed NetVault Servers: On multihomed NetVault Servers, configure the Preferred Network Address setting to allow the remote clients to send status messages to the correct address. For more information about this setting, see Configuring network addresses for multihomed machines.

If the remote clients fail to contact the server on the correct address, the task status is not updated on the server. In such cases, the clients are not added to the server and the task either remains in the Running state or completes with warnings.

•

Verify that the firewall is configured to allow traffic through the ports required for push installation: To push the client and plug-in packages to a remote Windows machine, the NetVault Server establishes the initial WMI session with the remote machine using RPC over port 135. All further WMI traffic uses the dynamic port range of 49152 through 65535.

To push the client and plug-in packages to a remote Linux machine, the NetVault Server establishes an SSH connection with the remote machine over port 22.

If there is a firewall between the server and the remote client, ensure that these ports are opened to allow the WMI RPC traffic to pass through the firewall. Also ensure that an inbound rule is created for the HTTP or HTTPS listen port that is configured on the NetVault Server.

When pushing the NetVault client installation to a Linux machine, you need the following additional prerequisites:

•

Grant the proper access: The user must have root-level access, with the NOPASSWD flag set in the sudoers file on the remote Linux machine. To complete this prerequisite, add one of the following entries to the sudoers file:

<username> ALL=NOPASSWD: ALL

<username> ALL=(ALL) NOPASSWD: <users home directory [/home/username or

/home/domainname/username]>/nvpushinstall/nvclientinstaller


	NOTE: If user fails to get sudoers privilege from the sudoers file, the deployment task will always be in the running state with Request sent to launch installation agent on the remote machine message.

•

Affirm the password authentication: In the sshd_config file, set the PasswordAuthentication entry to yes.

•

If you are not using the default cipher setting, then support for aes128-ctr should be present. If it is not, then you must add aes128-ctr at the end of the list following a comma in the etc/ssh/sshd_config file.

•

Install libstdc++: If you are pushing the hybrid installation package of the NetVault client on a 64-bit Linux machine, you must first install the 32-bit version of libstdc++.

When you push the NetVault client from a Linux machine to Windows machine, you need the following additional prerequisites:

•

Ensure the WinRM version is greater than 2.0.

•

Check WinRM on Windows client machine to verify Kerberos value:

Run the following command as an administrator in command line interface to open the config file.

winrm get winrm/config

Check the Kerberos value in the Auth section is set to true:

Kerberos = true

•

When using push install as a local user, ensure that the following WinRM configurations exist on the target Windows machine:

•