[lug] Thoughts on upgrading to CentOS 7

[Rob Nagler]

On Wed, Mar 28, 2018 at 6:43 AM, Alan Robertson <wrote:

> A complex solution to a simple problem that was already adequately solved.
>

>From a technical point of view, there are have been too many attempts at
solving this problem (including Sys V itself) in Unix for it to be
adequately solved. I find this interesting coming from a guy who works on
high availability stuff. Is there an article that argues sys v was a robust
solution to the process management problem? I would like to read it.


> But it's supporters were well-placed "politically" and their main
> arguments in favor of it were either easily solved in sysv (if anyone had
> cared), or were ad hominem arguments - variations on "You guys are all
> stupid" (replace stupid with luddite, etc).
>

What I don't understand is the statement that sysv easily solves the
process management problem. What in sysv restarts a service that has
failed? I can write this, and have, but that's not built into sysv afaik.
Operating systems need robust process management. Every sufficiently
complex Unix service ends up embedding it, e.g. Apache, Docker, and
Postfix. That's clearly been fixed by Systemd ("adequately" imiho). That
would be my main argument, at least. However, that's irrelevant as I'm a
user, and I don't get to choose, because I have other priorities.


> And if you want to argue that here, there's even less of a point now -
> it's been a fact for a while, and isn't going away for most people.
>

I generate all these files so it doesn't really matter to me one way or
another. The problem for me is the "bugs", which usually means poor or
incorrect documentation and defaults so I end up having to figure out the
"physics" of every new release. What does this parameter do? I dunno, it
says that it is the maximum number of frobs, but when I set it, it doesn't
do anything? Oh that's because it's controlled by this troggle. Rather
tricky, that.


> People like to say that technology wins in open source. This appears to be
> a counterexample.


Gerry Weinberg published the Psychology of Computer Programming in 1970. He
also later wrote "it's always a people problem", which gets quoted quite
frequently. Sadly, we rarely reflect on what our contribution is to the
problem. :)

This thread started as a lament: upgrading operating systems hasn't gotten
any simpler. The answer I realize is: people haven't gotten simpler. It's
what I come up with every time, and for some reason I always hope for
change. That's my contribution to my problem. :)

Rob
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lug.boulder.co.us/pipermail/lug/attachments/20180328/6e0b6a9f/attachment.html>