Skip to main content

No project description provided

Project description

Build status

pfm – a ssh port forward manager for data scientists

Data scientists launch a Jupyter Notebook servers to tackle each machine learning task. Usually local computers are not enough to handle multiple machine learning tasks. And therefore data scientists do their experiments in servers launched in remote hosts such as EC2 instances.

To connect Jupyter Notebook servers in remote hosts, we use ssh port forwarding. Port forwarding is useful since we do not consume resources in local PC.

Unfortunately, when connecting servers in multiple remote hosts and ports numbers, we easily forget the port number or assign the local port number which is used in another task. Especially when there are multiple remote hosts and ssh servers as the following image, understanding the combinations of remote hosts and local ports are difficult.

images/port-forwarding.jpeg

pfm manages the remote hosts and port numbers used in port forwarding. Users understand which local ports are used and which ports are not. Once users register the port forwarding information, pfm generates ssh parameters any time specifying the task name.

Install

We can install pfm with pip.

pip install pfm

Usage

Register settings of ssh port forwarding

pfm add registers port forward settings.

Register settings with options

The following is a sample of adding port forward settings with command line options.

$ pfm add -n image-classification --local-port 9999 --remote-port 8888 --ssh-server myml.aws.com --remote-host localhost

pfm add provides the following options.

Usage: pfm add [OPTIONS]

add port forwarding target

Options:
    -n, --name TEXT        name of port fowarding
    --forward-type TEXT    port forwarding type [L (local) or R (remote)]
    --local-port INTEGER   local port
    --remote-port INTEGER  remote host port
    --ssh-server TEXT      server to ssh login
    --server-port INTEGER  server port
    --remote-host TEXT     remote host for port forwarding
    --login-user TEXT      login user of ssh server
    --help                 Show this message and exit.

Note that when local port number (--local-port) is not specified, pfm add automatically assigns the local port not to collide to other port forward settings.

Register settings with a argument

If you think adding many options are tedious, pfm add also provides argument style of registration.

The following is a sample of registration which forwards port 8888 in takahi-i-ml.aws.com to port 9999 of localhost with a local port forward setting.

$pfm add -n image-classification "9999:localhost:8888 takahi-i-ml.aws.com"

As we see the above example is simple and just uses -n option. Note that default forward type is set to local (L). If you want to set forward type to remote please specify the type with --forward_type parameter.

Generate ssh port forward parameters

After the registration of port forward settings with pfm add , we can generate ssh parameters with pfm param.

$ ssh `pfm param image-classification`

Or we can also use pfm connect in order to do same (Note that this command includes -A ssh option).

$ pfm connect image-classification

Update port forward settings

After the registration of port forward settings with pfm add , we can modify ssh parameters with pfm update. For example, the following command changes the port forwarding type of image-classification setting from local to remote.

$ pfm update --name image-classification --forward-type R

List registered ssh port forward settings

We can see the list of registered port forward settings.

$ pfm list
+----------------------+------------+------------+--------------------------------+--------------+-----------------+--------------------------------+--------------+
|         name         |    type    | local_port |          remote_host           | remote_port  |   login_user    |           ssh_server           | server_port  |
+======================+============+============+================================+==============+=================+================================+==============+
| image-processing     | L          | 9999       | localhost                      | 8888         | None            | my-ml-instance.aws.com         |              |
+----------------------+------------+------------+--------------------------------+--------------+-----------------+--------------------------------+--------------+
| text-processing      | L          | 7777       | localhost                      | 8888         | None            | my-ml-instance-2.aws.com       |              |
+----------------------+------------+------------+--------------------------------+--------------+-----------------+--------------------------------+--------------+

Delete registered forwarding setting

When a port forward settings is not needed, we can remove the setting with pfm delete command

$ pfm delete image-process

License

  • Free software: MIT license

Resources

Slides on pfm.

For developers

We are welcome any contribution. For details, please see CONTRIBUTING.rst.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for pfm, version 0.6.0
Filename, size File type Python version Upload date Hashes
Filename, size pfm-0.6.0-py3-none-any.whl (11.0 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size pfm-0.6.0.tar.gz (10.0 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page