What is our primary use case?
The use case in my very early years was for dedicated servers for doing web applications.
How has it helped my organization?
We almost exclusively use Red Hat. The benefits boil down to the support. There is no problem getting support. Whenever we have an issue that we cannot solve, which does not happen often, we have somebody who is there either virtually or physically.
We use Red Hat Enterprise Linux on-prem and on the cloud in a hybrid environment. We probably also have edge devices. I am not completely sure about that one. Having it in a hybrid cloud deployment has been no different than having it on-prem. Running it on-prem is just as good as running it on the cloud for us. It simply works.
I appreciate the dashboards that are available online. There has been a lot of feedback on the CVEs. The most recent one that came was probably related to Zutil. Red Hat made an announcement very quickly saying that if you are using only Red Hat features, you do not have to worry about it. It does not run on their operating system. Unless you are custom compiling, it does not work on their system. I greatly appreciate little things like that because they save us a lot of time. If Red Hat is simply saying that it is not a part of their repo, I do not have to look for it.
We use Red Hat Insights but not company-wide. It is one of those things that simply saves you time. I do not want to have myself or anyone on my team go out and check various things. That is the whole purpose of using Red Hat Satellite. The whole purpose of all different dashboards and these websites is to use what you have. Let it report out what you have and not continue to write scripts just to check things.
What is most valuable?
Their support is valuable. Whenever I had a problem, I could get on a phone call with somebody. I did not have to go to some random forum or send an email and wait forever. I could call somebody.
What needs improvement?
It does have a workstation option, but you rarely hear anything about it. I would love to see the workstation replace Windows. That is a stretch goal, but it is possible.
For how long have I used the solution?
I have been using Red Hat Enterprise Linux since version 4. It has been a while.
What do I think about the stability of the solution?
It is stable as long as you do not do something stupid.
What do I think about the scalability of the solution?
Red Hat specifically works hard to make it difficult to not be able to scale it into anything. The only thing that I do not see it being capable of, officially at least, are the IoT devices. Technically, it is possible to get it on those devices, but that is not something Red Hat is focusing on right now. From a scalability standpoint, it comes down to what makes a reasonable profit and what is a good return on investment while choosing how to scale and where to scale. Red Hat is doing it right so far.
How are customer service and support?
Prior to a few months ago, the support that we got from a TAM point of view was next to none. Now that I understand the scenario a little bit more, it was not because Red Hat was not doing its job or did not want to do more support. It was because of how the contracts aligned, and more importantly, who in our organization was handling those contracts. We had a recent change in our organization in terms of who is running what and who is handling what. When that change happened, the doors really burst open. Now that we have a different person he is working with, we are getting incredible support from our TAM. He is in communication with us on a very regular basis. While I have been here at Red Hat Summit, we have gone out to have meetings twice. I cannot speak highly enough. I would rate their support a ten out of ten.
How would you rate customer service and support?
Which solution did I use previously and why did I switch?
My current organization has pretty much always used Red Hat, specifically Red Hat Enterprise Linux. There are all sorts of flavors of Unix in our environment. Almost all of them are there because they are managed network devices.
We wanted to stay close to Red Hat Enterprise Linux simply because of the mentality of the business. We have got some people who have been around for 20 years. Things such as switching from YUM update to APT update are easy. People can usually change from one to another pretty quickly, but some of the other commands that you are used to running in Red Hat Enterprise Linux are slightly different for different versions of Unix. It did not make sense.
I have used a lot of different variants through the years. I could be running Raspberry Pi, or I could be using Ubuntu to do a job but not for the production environment. I do not waste my time anymore. I know what works and where support is.
How was the initial setup?
Our setup is a bit of a hybrid. We are streamlining a lot of things and trying to redesign how we are doing things. In terms of the cloud, we are 100% TerraForm. We are building out infrastructure as a code and TerraForm pipelines. On-prem, we have a Jenkins job that runs some TerraForm, which then runs some Ansible and then some Puppet. There is some cleaning up needed there.
Currently, we use all three major cloud providers: Azure, Google Cloud, and AWS. Each has its purpose.
The initial experience of deploying it at the current company was terrible, but it was not a Red Hat issue. It was an internalized issue that took a little bit of time to work out. After that, it was not a problem.
What about the implementation team?
We implement it on our own.
What was our ROI?
I have not run into a single person who knows about Red Hat Enterprise Linux and is not being helpful. You can get talking with somebody at Red Hat Summit about what you are doing on Red Hat Enterprise Linux, and they will be like, "I did that a couple of days ago. Did you run into this problem too?" There is a community. I am sure there are communities for other variants, but my return on investment is simply community and support. I cannot speak highly enough of these two.
Which other solutions did I evaluate?
To a colleague who is looking at open-source, cloud-based operating systems for Linux instead of Red Hat Enterprise Linux, I would say, "Good Luck!" We looked at a lot of different options to potentially leave Red Hat simply because of the cost. We tried out CentOS. We tried out Rocky. There were even talks about trying out Ubuntu, but there was the hassle of changing all of our mentality and code to work with different systems. It just did not make sense. CentOS worked almost side by side with Red Hat, but certain things that we have specialized with Red Hat were not working on CentOS for some reason.
We chose not to use CentOS because we had a misunderstanding of what AppStream was in terms of end-of-life for CentOS. Rocky was ruled out pretty quickly simply because of a lack of understanding in terms of:
- Where does Rocky come from?
- How reliable is it?
- Where is the support?
Red Hat's support model trumps a lot of those other ideas. I tell people that even if they are working in a home lab environment, get a developer license and get a developer account with Red Hat. Use Red Hat because more and more businesses I work with simply use Red Hat. It is great to have Fedora on your laptop as a workstation. It is great to have CentOS as a workstation. That is because those are still a part of Red Hat. You can transition and use Red Hat for a company. I have not been a fan of Ubuntu and some of the other variants because of how easy it is for people to make changes to operating systems that are not fully backed or tested. In my opinion, you do not want to put production on it.
What other advice do I have?
Red Hat Enterprise Linux has not enabled us to centralize development. We are moving towards centralized development, but there are still so many different teams, so centralized development is not yet there.
We are partially using Red Hat Enterprise Linux for containerization projects. Within the next year, I hope to bring OpenShift in and replace AKS. I do not have a use case for the portability of applications and containers built on Red Hat Enterprise Linux. Based on what I have seen here at Red Hat Summit, I have a lot of ideas spinning around in my head to make it happen, but I do not yet have anything around containerization.
Red Hat Insights provides vulnerability alerts and targeted guidance, but we are currently not using that side of it. It helps in my limited sandbox environment, but of course, my sandbox is built up and torn down like crazy. It is valuable, but we do not have a great use case yet.
I would rate Red Hat Enterprise Linux a ten out of ten. I have been working with Unix systems for a while now. The first Unix system I touched was in 1992. There were so many variants that were striving to become well-known. You would hear all of these weird names. There were all of these weird animals and all of these different logos through the years. Even before 1992, there were a lot. As things progressed, you quickly saw different ones die out. I do not remember when I truly got onboarded with Red Hat. I know I started with version 4. It is one of those companies when you are looking for a name that sticks around and about which you do not have to question if they are going to be around for a while. You do not have to question that with Red Hat. You do not have to question that with Red Hat Enterprise Linux, whereas a lot of other variants do not even exist anymore, or they exist, but they have not been maintained longer than some people have been alive.
*Disclosure: I am a real user, and this review is based on my own experience and opinions.