Hi all -
I have some difficult people in my organization that are in positions of power that think that just because doing something manually “isn’t very hard”, it shouldn’t be automated. For example, clone of VMs from templates, setting up sshd_config, etc. Is there anyone out there that would be willing to exchange emails (so I’m not airing dirty laundry publicly) to trade ideas on how to deal with these difficult situations? Bonus if you’re in/near the DC or Virginia, USA area.
Thanks,
Adam
Hi all -
I have some difficult people in my organization that are in positions of power that think that just because doing something manually “isn’t very hard”, it shouldn’t be automated. For example, clone of VMs from templates, setting up sshd_config, etc.
Hm… such people should not be in power - perhaps if we knew more about their reasoning for this attitude then the list may be useful in proposing lines of arguments to convince them?
Is there anyone out there that would be willing to exchange emails (so I’m not airing dirty laundry publicly) to trade ideas on how to deal with these difficult situations? Bonus if you’re in/near the DC or Virginia, USA area.
By the little you have revealed, I doubt they would be receptive from advice from a random stranger 
--
Karl E. Jorgensen <jorginator74@gmail.com>
I agree, they shouldn’t be in power. Alas, it’s the situation I’m in, and I’m not a Director/VP. The people I refer to are system engineers. My thought is they are comfortable being the only people with root access (even though I’ve managed to get Ansible setup with some level of root privileges), and see it as a job security threat to allow other people such access. The other “reason” I’ve received is that any mature organization has “division of duties” to ensure no one person has too much power. Which is true, but I’m trying to ensure business continuity without having to remember (and make manual changes) to systems that have special customizations. I’ve tried to explain that as the platform & business owner, I’m the one that has to answer when things go wrong. I also have a proven track record of including people in decisions, but it seems like a control thing. I just want consistent, reliable, repeatable upgrades.
Adam
I’d suggest automating stuff for any environments you have control over (say development or just virtualization running on your machine if necessary).
You can give other teams the ability to control password access (probably lots of ways to do this - keep inventory in separate repo and let other teams manage vault passwords, for example).
If its automated you can at least make use of that and there’s code ready to go when it makes sense for other teams to use it. I guess I’m saying go for a ‘lead by example’ approach. Its always better when people can choose change and showing tangible benefits is going to make it easier to choose to change.
Otherwise I’d suggest you look for another gig, sounds like your current organisation is taking the wind out of your sails.
Hope this helps,
Jon
This is currently an opinion argument, and you need to turn it into a
numbers argument. It's the only way to get past egos.
If a team deliver what they're asked to do on spec and on time, I
personally don't care what tooling they use, whether it's ansible or
hand-crafted.
If they're delivering late or low quality, start quantifying that and
maybe trial some automation, then compare the outputs.
Thanks to everyone for the feedback. This was all very useful!
-Adam