pcp
[Top] [All Lists]

Re: [pcp] [RFC] Minimizing Installation Size for Reduced PCP Footprint

To: Mark Goodwin <goodwinos@xxxxxxxxx>
Subject: Re: [pcp] [RFC] Minimizing Installation Size for Reduced PCP Footprint
From: Nathan Scott <nathans@xxxxxxxxxx>
Date: Wed, 6 May 2015 03:08:10 -0400 (EDT)
Cc: Lukas Berk <lberk@xxxxxxxxxx>, pcp@xxxxxxxxxxx
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <5549901B.60809@xxxxxxxxx>
References: <87bnk0wzn5.fsf@xxxxxxxxxx> <1565443824.10833179.1430459606977.JavaMail.zimbra@xxxxxxxxxx> <87383c19uy.fsf@xxxxxxxxxx> <554843FC.9040109@xxxxxxxxxx> <87d22eyfut.fsf@xxxxxxxxxx> <55496FC7.8060600@xxxxxxxxx> <1048697364.13608682.1430878823251.JavaMail.zimbra@xxxxxxxxxx> <5549901B.60809@xxxxxxxxx>
Reply-to: Nathan Scott <nathans@xxxxxxxxxx>
Thread-index: EVr6+/RxWZ//ty7e2iJEI3Fb0O4SNA==
Thread-topic: Minimizing Installation Size for Reduced PCP Footprint

----- Original Message -----
> On 05/06/2015 12:20 PM, Nathan Scott wrote:
> >
> >
> > ----- Original Message -----
> >> Looking pretty good now - we're getting close :) More comments below ..
> >> [...]
> >
> > +1
> >
> >> OK, so now pmiostat has been pcp-foo'erized but there is no pmiostat
> >> symlink. Also the man page has disappeared after the upgrade (see below).
> >
> > This is accidental I think, there must be at least a symlink for
> > back-compat.
> 
> yep. I'll look at a patch for this later today if I get time ..

(done in my tree)

> >
> >> [...] but my question above
> >> still stands - what are people's thoughts on this? I personally don't
> >> really like it, which is why I originally shipped pmiostat rather
> >> than pcp-iostat (and also because pmiostat isn't an exact drop in
> >> replacement for iostat .. the latter has too many warts and the args
> >> can never match up aynway) - I'm a complete and utter anti-sysstat
> >> bigot!
> >
> > :)  I think the realities of providing drop-in replacements mean its a
> > losing battle chasing perfect exactness, and look-alikes will have to do.
> 
> agree, it's definitely a loosing battle (from the args usage POV alone)
> 
> > re install location, the intention was always to let people choose, cos
> > different people have different needs/ideas.  For me - over time I hope
> > to build up the front end tools we have with more look-alikes, and the
> > thought of so many pm* tools in the system PATH /usr/bin gives me pause.
> > Also, if we were to enforce pm*, many will involve swallowing your own
> > tongue trying to pronounce the names (like pmmpstat - pcp-mpstat is alot
> > clearer).
> 
> yes certainly agree pcp-foo is clearer. And I especially like autocompleting
> pcp-TABTAB to see all the pcp-foo tools available. I'd like to see the
> pcp-foo tools elevated to /usr/bin where we don't need an augmented $PATH
> rather than hiding in /usr/libexec... So how about we move them to /usr/bin
> and also, when appropriate, ship a symlink for pmfoo (for backcompat).
> The man pages can be aliased too.

I prefer the git model, where there sub-commands are not splattered all
over the PATH (still get command completion though - doesn't need to be
on the path for that).

> > There's pros and cons both ways, and we can get best of both
> > with the symlink approach when need be.
> 
> agree
> 
> and whilst we're at it, why are pmlogconf and pmlogger hiding in libexec?
> Especially pmlogconf since it has an interactive mode, though I don't
> know anyone who actually uses that.

$ rpm -qf /usr/bin/pmlogger
pcp-3.10.5-1.x86_64

(and theres a compat symlink in PCP_BINADM_DIR - could do the same for
logconf if its wanted on default PATH?)

cheers.

--
Nathan

<Prev in Thread] Current Thread [Next in Thread>