Introduction to the community

Hello. My name is Gary Scarborough and I work for Red Hat as a Support Engineer in System Management.

I have 2 ideas currently for articles.

  1. Common mistakes in automation

  2. Introduction to VDO

Thank you,

Gary

17 Likes

Welcome Gary. Automation and Virtual Data Optimizer are interesting topics. More detail on the type of automation would be useful.

1 Like

hello!

Hello Garry,
Welcome to the community. Looking forward to seeing your articles on the magazine.

Jou, Garry. Greetings from Eu.

Yes, welcome, Gary! I second @ fed500 …both common automation mistakes and VDO would be welcome topics!

Welcome aboard, Gary!

Welcome, those topics would be awesome :slight_smile:

Hi Gary,
Not that I spent much time on this, but here is there I would focus.

  1. Common mistakes in automation

    The biggest mistake is, without doubt, company’s not spending the time and money on the Planning, Design, Analysis, and Maintenance of the correct standards, of the required automated system or functionality required…

They try to use something like standard Fedora Linux as a cheap real-time alternative, or use open network devices as a base, etc,…

Requirements that creep all over the place!

Fedora is a good, very good, starting point, as a Development platform, but you must be able to stripe out all the unnecessary Fedora systems. e.g GUI/Gnome, Audit, logging, Security (I’m not saying you don’t need Logging or Security, etc,… just different or smaller type).

  1. Introduction to VDO

This is very similar to system design problems above; Requirements that creep!

For example; Designing a GUI specifically for the device. So, removing Gnome, etc,… And using/finding a open source library suitable for GUI (Not using C++ because it’s too large and not C).

Given that I’m not a front end (GUI) or high level (App) developer (I do not use C++, or the like. It’s a pain to have to stop and change tool chains and find relevant library’s), I spend far to much time looking to remove or stripe a distro like Fedora (This is just for a Development platform). I tend to start with a base Kernel, and add/configure what is needed, for the device, itself.

Company’s are refusing to use something like Fedora or some distro (open source), because of past mistakes, difficulty of maintaining the platform. Things like the Video driver on my Fedora 36. I have not been able to get my Intel video device running, or run MS Teams on it.

Hope this helps?

Regards,

Roger Moore.

Hello!

automation errors keep happening; but who detects them ?

automation errors keep happening;

but who detects them ?

and corrects them before damage takes place?

Hello,

The topic applies to all forms of automation. These are common issues I see in infrastructures both large and small. The 4 main issues presented by the article are:

1.scripts are not tested thoroughly before deployment
2. admins do not consider the load the scripts will place on resources
3. admins forget make changes that are then reversed by compliance scripts
4. admins do not document their work

I am part way through a rough draft.

1 Like

Placing too much responsibility on admins is not automation. )

The biggest mistake is, without doubt, company’s not spending the time and money on the Planning, Design, Analysis, and Maintenance of the correct standards, of the required automated system or functionality required…

I would say the biggest mistake is the absence of peer reviewed automation scenarios with diagrams to choose from. For example, the simplest scenario.

How do you keep Fedora instances updated?

  • Do you login manually to execute dnf update and review the changes?
  • Do you reboot immediately after update? Or do you have a scheduled time for reboots?
  • Do you get an email about urgent security updates that need to be installed ASAP?
  • Do you you have everything above automated, so that you only get notification reports?
  • Have you managed to achieve zero downtime while updating?

What, from a Peer with a PHd in Biology ?

Not that I accept a PHd as a good basis for an Engineer!

The responsibility rests on the company, and it’s staff, doing the project.

If you don’t have a Good Engineering manager (With a relevant technical degree), and a methodology to match, you will end up all over the place!

Peer review would be meaningless without it!

I haven’t heard that scientists employ Scientific managers to do peer reviews, so I would say that quality of peer review is a property of company culture, and not a product of hierarchy and hiring luck.

What Rubbish!

Who said anything about a Science Manager?

Peer review, in on it’s self, is not suitable for an Engineering environment, neither is Science more generally.

Company’s have to employ a capable Engineering Manager, with a suitable Methodology (this includes suitably experienced Engineers)!

So what do you propose for Fedora Project?

Why would you ask me that?

How do I know how the Fedora Project is managed?

But, I would suggest that even for a community project, there has to be management, and standards for, the parcels, before you can put them altogether ?!

And, that still wouldn’t be as a Science project, but an Engineering program.

I didn’t mean to confuse anyone, I was speaking from a commercial perspective.

So, if you don’t have an Engineering Manager, you can still manage it well.

But, do what works, and in time/experience, you will learn what works for this type of development.

Develop your own methodology, and maintain your own standards?

Etc,…

Well, because it is the Fedora Project.

Read, ask, express interest? I thought that a person with strong opinion would be naturally be interested in proving their theory in the field. And as a new member of the community would be pleased to review and improve existing practices.