Add command for tailing logs about cage HOT 7 CLOSED

Could we spec out some ideas about how this might work in an ideal world? We don't have to implement all right away, but I'd like to get the API right.

Here's my proposal for the final API:

cage logs [log options] [<pods_or_services>...]

This covers a number of cases:

• If it displays logs for more than one container, it annotes each line with the container name.
• With no args, it displays interleaved logs for all containers.
• With a pod name, it displays the logs for all containers in that pod.
• With either pod/service or a unique service name, it displays the logs for that service.

from cage.

The standard Docker tools include the following options:

Usage:  docker logs [OPTIONS] CONTAINER

Fetch the logs of a container

Options:
      --details        Show extra details provided to logs
  -f, --follow         Follow log output
      --help           Print usage
      --since string   Show logs since timestamp
      --tail string    Number of lines to show from the end of the logs (default "all")
  -t, --timestamps     Show timestamps

docker-compose logs --help
View output from containers.

Usage: logs [options] [SERVICE...]

Options:
    --no-color          Produce monochrome output.
    -f, --follow        Follow log output.
    -t, --timestamps    Show timestamps.
    --tail="all"        Number of lines to show from the end of the logs
                        for each container.

from cage.

@dkastner Let me fix #8 first (the new argument-parsing scheme), and then you can go ahead and implement any subset of this that you want to work on.

from cage.

The fancier, multi-container versions of this might benefit from #9, which would make it easier to multiplex tagged I/O from multiple threads or external processes. But that doesn't need to block the initial implementation.

from cage.

OK, I've added a project.pod_or_service(name) function which can be used to look up either a pod or a service to print the logs from. This will allow you to easily implement:

cage logs <POD_OR_SERVICE>

I wouldn't worry about the multi-pod / multi-service version of this command yet, because doing a good job of that depends on #9, which I want to poke at some more.

To implement this, I would recommend doing the following:

1. Edit src/cli.yml to add a new logs subcommand with appropriate arguments. See the clap documentation.
2. Dispatch the new arguments in src/main.rs to your new src/cmd/logs.rs. This should look like:
   rust pub trait CommandLogs { /// Up all the images in the specified pods. fn logs<CR>(&self, runner: &CR, ovr: &Override, pod_or_service_name: &str) -> Result<()> where CR: CommandRunner; }
3. From inside that function, call match try!(self.pod_or_service_or_err(pod_or_service_name)) { ... } and handle both the PodOrService::Pod and the PodOrService::Service case.

from cage.

Also, if you want to pass through any arguments, you can add a new args::opts::Logs structure to src/args/opts.rs, based on the ones already there (but ignore the Deref hack; that's going away shortly) and implement ToArgs and add method to ArgMatchesExt in src/main.rs that builds your args::opts::Logs structure based on the ArgMatches structure from clap.

If you have any questions, please let me know!

from cage.

There are also new goodies in 488bf19 which might be of interest. Although you probably can't use the compose command directly, it would be easy to make a single-item wrapper around it.

from cage.