[talk] OpsMops Halted (probably OT but tangentially related to April 3 talk)

Edward Capriolo edlinuxguru at gmail.com
Mon Apr 1 21:59:53 EDT 2019

On Mon, Apr 1, 2019 at 9:19 PM Raul Cuza <raulcuza at gmail.com> wrote:

> (1) https://threadreaderapp.com/thread/1091710068234641408.html
> (2) https://news.ycombinator.com/item?id=18716628
> (3)
> https://techbeacon.com/devops/how-create-devops-testing-culture-3-keys-success
> # tl;dr
> (1) is the announcement that Michael DeHaan is stopping development of
> OpsMop only after a few months of working on it. This is the same
> developer who wrote Ansible.
> (2) is a small selection of the reaction to DeHaan's announcement
> about OpsMop and proof that his reasons for stopping development have
> merit. (More about the reasons for his stopping below.)
> (3) is a tech BS jargon heavy article that shows that there are other
> people who agree with DeHaan but have their head too far up their
> chosen solution to see the solution to their misery.
> # Why am I sharing this?
> I miss the (nonexistent?) tech era when the hardware, operating system
> and application were all crafted to work together by the same team;
> choices were optimized for the people, problem and situation. Building
> out a tech solution today involves so many layers piled on each other
> that it is a miserable experience troubleshooting through all the
> garbage code. Maybe if I had some choice over which open source OS I
> had to run on other people's hardware, I would not sound like a grumpy
> old sysadmin right now.
> I sympathize with DeHaan who wishes DevOps didn't mean DSL writing
> SysAdmins who work only on ops and automation. I also don't want to
> fall into his trap of having to be "feed by interaction" which is
> ultimately why he stopped work on OpsMop. If a project scratches an
> itch, then why not work on it whether you are going to be on the cover
> of PC Magazine or not? But then again, I am not a tech artist, so
> maybe I don't understand.
> Finally, Adam Auerbach in "How to Create a a DevOps Testing Culture"
> essentially agrees with DeHaan, but for the sake of DevOps Culture™
> which misses the point.
> """
> *Successful DevOps adoption* requires more than just automation across
> the software development lifecycle; you also must change the way your
> team is organized, the way you work, and the expectations you have for
> people on your team. It means having engineers in all roles and
> removing the dependency on enterprise IT teams—such as operations or
> security testing—from getting features out the door.
> """ (my emphasis)
> The point isn't DevOps adoption, but building the effective tech
> solution efficiently. I do agree that all the Go/No Go people should
> be on the same team and working together. Having outside people who
> don't take the time to understand the technology but can stop its
> deployment is a sad situation. Gatekeepers are feudal "Expert Class"
> that should be democratized by spreading the skill around evenly (i.e.
> turn their expertise into testing code and engineer can read!).
> Raúl "really needs to stop procrastinating" Cuza
> p.s. Maybe I should be more forgiving as in my day I wielded tech
> jargon like I had found the Holy Grail. But I'm older than that now;
> so meh.
> _______________________________________________
> talk mailing list
> talk at lists.nycbug.org
> http://lists.nycbug.org:8080/mailman/listinfo/talk

"We get tired, and we let things just go. We use what other people are
using. We run clouds on top of our freaking clouds for no reason, and we
are much more interested in tech fashion than what makes us productive. We
tolerate software with thousands of bugs."

Really? but Ansible which wraps python on top of shell commands is more
productive? I used some of the first versions of ansible.They were bad, in
particular the entire "this sexy python thing does not work with RHEL,
because of strange python ssh library and having to patch ssh server to
involve pipelining", and it "worked" out of the box but ran as slow as dirt
because it would uses some python math module instead of c-code to do ssl
:) Touching a file was wrapped in multiple levels of python, ow and the
YAML files... And out for every command I wanted to run it was running 19
other commands to "fact check" my system.

Most programers don't know much about operating systems or networking, most
ops people hate writing code. All dev-ops is now is now just what every
sysadmin calls themselves when getting a new job. If you reject the devops
labelgo going even more cutting edge and call yourself an SRE. lol

Open source software feeds off interaction. And this is my diagnosis
basically - widescale burnout. It's been coming in slow but we didn't see
it. Agile and DevOps are the burnout. IMHO we are *poorer* technologically
than we were 6 years ago in some areas. Why

Not really what is going on. In the 60s people theorized that computing
would soon enter the conceptual age. IE I tell my compiler "make me a
program with 2 text boxes and a button that when I click on the text boxes
it adds them".

If a AI program can write code, why should a human have to write code to
setup a system.Think like back to the future 2....

"You have to use your hands....thats like a baby toy"
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.nycbug.org:8080/pipermail/talk/attachments/20190401/ade33eae/attachment.html>

More information about the talk mailing list