Execute command on Kubernetes node
Execute a command on a node.
This executes a command directly on the node itself. Therefore, havener creates a temporary pod which enables the user to access the shell of the node. The pod is deleted automatically afterwards.
The command can be omitted which will result in the default command: /bin/sh. For example 'havener node-exec foo' will search for a node named 'foo' and open a shell if found.
Typically, the TTY flag does have to be specified. By definition, if one one target node is provided, it is assumed that TTY is desired and STDIN is attached to the remote process. Analog, for the distributed mode with multiple nodes, no TTY is set and the STDIN is multiplexed into each remote process.
If you run the 'node-exec' without any additional arguments, it will print a list of available nodes.
For convenenience, if the target node name all is used, havener will look up all nodes automatically.
havener node-exec [flags] [<node>[,<node>,...]] [<command>]
--block show distributed shell output as block for each node
-h, --help help for node-exec
--image string set image for helper pod from which the root-shell is accessed (default "alpine")
--no-tty do not allocate pseudo-terminal for command execution
--timeout int set timout in seconds for the setup of the helper pod (default 10)
--debug debug output - level 5
--error error output - level 2
--fatal fatal output - level 1
--kubeconfig string Kubernetes configuration file (default "~/.kube/config")
--terminal-height int disable autodetection and specify an explicit terminal height (default -1)
--terminal-width int disable autodetection and specify an explicit terminal width (default -1)
--trace trace output - level 6
-v, --verbose verbose output - level 4
--warn warn output - level 3
- havener - Convenience wrapper around both kubectl and helm