Jan,

Currently the 'pcs <...> -h' should only be giving details on the command you're trying to run.

For Example:

[root@rh7-2 ~]# pcs status pcsd -h

Usage: pcs status pcsd...
    pcsd <node> ...
        Show the current status of pcsd on the specified nodes


It doesn't show the full status output, only the instructions for 'pcs status pcsd'.

Or here, where there are two possible commands:
[root@rh7-2 ~]# pcs resource failcount -h

Usage: pcs resource failcount...
    failcount show <resource id> [node]
        Show current failcount for specified resource from all nodes or
        only on specified node

    failcount reset <resource id> [node]
        Reset failcount for specified resource on all nodes or only on
        specified node.  This tells the cluster to forget how many times
        a resource has failed in the past.  This may allow the resource to
        be started or moved to a more preferred location.


It doesn't show all 'pcs resource' commands, just the ones that start with 'pcs resource failcount'.

While it isn't as compact as the ip commands, it does provide the full information for the user without using too many lines of output.

I'm marking this at WONTFIX (for now) because I don't think it makes sense to implement this in the near term.  If we do get lots of complaints, we can always re-open and revisit this issue.

Fair enough.  But from what I've heard from Tomáš, there is long-term goal
to have single authoritative meta-source for all help screens, man pages
and bash completion, so if there's ever and overhaul in this area, please
reconsider this bug (keeping the respective data in a structured way is
3/4 of the solution for this bug, anyway).

Definitely. Once we have one place for all the documentation this will be a lot simpler to implement/maintain.

Also, one can anticipate a popular demand for such a reference, as
demonstrated by hand-crafting requested overview just to point out
a missing item on the ML:

http://clusterlabs.org/pipermail/users/2015-April/000251.html

re [comment 0]:

> 1. add "help" subcommand, where "pcs help SOMETHING" would be a direct
>    equivalent of "pcs SOMETHING -h" (incl. SOMETHING being left out)
>    - this is quite popular, used in git (and plenty of other VCSs),
>      virsh, etc. and can pretty much be first choice for *nix CLI
>      newcomers

Another (even Red Hat associated) pieces of software supporting "help"
subcommands: yum and dnf.

I am reopening the bug as it needs new reconsideration and either closing
as duplicate to wider-scoped bug (context including bash-completion, etc.,
see [comment 2]) or properly justifying why the desired high-level
overview is not provided (more on this below).

Note that for the purpose of clufter (colorozing the longest valid prefix
sequence of command atoms in pcs syntax), I've manually tracked down
the overall commands hierarchy:

https://pagure.io/clufter/blob/b4b12614e1d243c0ed21b3be7896a153fb02d219/f/filters/cmd_wrap.py#_23

and guess what ... there are some 180 basic, standalone pcs commands
(not counting the possible variations of each).

In this light, I consider it a must to have a syntax reference as
requested (overall plus possibility to drill down with command atoms),
only to allow one to grok the overall scope of what can be done.
(Utterly) verbose/long help screen won't cut it for this purpose.

It can be speculated that people miss such basic commands as
"pcs cluster node add" because of the lack of requested syntax
reference/commands overview:

http://oss.clusterlabs.org/pipermail/users/2016-August/003715.html

After evaluating this issue, there are no plans to address it further or fix it in an upcoming release.  Therefore, it is being closed.  If plans change such that this issue will be fixed in an upcoming release, then the bug can be reopened.