What is our primary use case?
We run web apps. We run databases. We run a high-compute platform on Red Hat Enterprise Linux variants.
All of our customers run Red Hat Enterprise Linux. We run Red Hat Enterprise Linux for mesh nodes. For anything Linux, if we can use Red Hat Enterprise Linux because it is supported, we put it on Red Hat Enterprise Linux. Probably 60% to 80% of our infrastructure is Red Hat.
How has it helped my organization?
Having a stable Linux platform means I am not spending my time rebuilding Linux systems, constantly patching, and doing things like that. It helps to have an approved and supported platform. I know they have tested everything and when I patch my system, it is not going to blow up. It just does not happen. The other thing is that we have had catastrophic failures, and they have helped us out of these catastrophic failures. The support for Red Hat Enterprise Linux has always been good, and the community around Red Hat Enterprise Linux has been fantastic.
We were also CentOS users, so we have committed to AppStream as well. Being a part of the community has been a huge benefit for us. Community adoption means it is easy for people to find information. It helps new people get on boarded into Linux.
We mostly have an on-prem environment. VMware is a significant chunk. We do have some Red Hat clusters. We do have clustered applications, both physical and virtual, running on the cluster. We do have some cloud. We have our own internal cloud with VMware running behind the scenes. Having a consistent image means things always look the same. It is boring, but it is cookie-cutter. That is what we like. We like everything to come out the same. We have consistency and the ability to patch across our entire environment. We are also a Satellite user, so we are able to patch everything and maintain everything in a single pane of glass. It means I can have fewer admins administering many more machines. If you have a reduction in failure and an improvement in automation, things just work.
We have created what we call creator nodes. We have built a platform on Red Hat with Podman so that they can connect with Visual Studio code and do development or Ansible development. We now have our mainframe people developing automation with Linux with all of the plugins right there. It is a consistent environment for them, and that has been awesome. That has been fantastic. We have a few hiccups with Podman. They are working on the permissions to be able to have multiple people run Podman. They are working on the UID and GID problem that we had earlier. Right now, we are running Docker, but I am planning on moving to Podman once they fix that. We have also automated the build process for those nodes. If we need to scale up, we build a couple more VMs, and we are done.
We use Red Hat Enterprise Linux for containerization projects. We are containerizing applications. We are pulling the Windows container that we have and converting it to the Red Hat Enterprise Linux container. At the Red Hat Summit, the keynote about image RHEL with systemd blew my mind. It is a change from what we have been doing, but it should make a lot of things more reachable for us. It is cool because now my container image looks like my VM image. You cannot make it simpler for people to develop in a container. It looks the same. There is no difference. That is going to drive heavy adoption with us because if there is no difference, people are not going to have that fear of something new. It has 100% impacted our projects in a positive way. We have started to migrate all of our workloads to OpenShift now that we have got it in the door. It makes a lot of sense. I can redeploy. I can patch. I can do all this with code. I do not have to maintain a VM and a container. It makes life simple.
We have seen a drop in TCO because we ended up buying more than building. When you build something, there is the hidden cost of support, training, and the precarious position you get in if you deploy something you do not fully understand. We were there. We had five instances and a bunch of complexity. We reduced that down to one. We were able to simplify our complex nature. That is what Red Hat has allowed us to do. We have been able to roll out and we have been consistent. I have got machines out there that have been running for two or three years with no problems. They just patch them in the background. It just works.
What is most valuable?
I love systemd. They have made some significant improvements with the firewalld console. I do not use it that much, but I know it makes Linux reachable for people who are not normally Linux admins.
I just love the command line configuration. It makes that easy for me. Another thing is that when you combine that with Ansible, your life is simple. You can do a lot of your jobs without having to touch the system. That is my ideal.
I appreciate everything they have done. The systems are just bulletproof. We do not have problems with it. Support for file system differences and migrations has been solid.
What needs improvement?
There have been a few things that I have run into. They have significantly improved DNF and YUM, but there can be better communication around what is going on. A lot of it is related to communication. They are building solid products, and quite often, people do not find out about them until two or three years have passed. We still have not discovered everything in Red Hat Enterprise Linux 9. A lot of it is because we have not had the time, but it would be helpful to have a little bit more communication around it. Maybe that is on us to make sure that we stay updated with the community.
For how long have I used the solution?
I have been using it since Red Hat Enterprise Linux 2.5. It has been around 20 years. I love Red Hat Enterprise Linux.
What do I think about the stability of the solution?
I would rate it a nine out of ten for stability. It is stable. It is fairly bulletproof. There are a lot more things that they are adding to make it better.
What do I think about the scalability of the solution?
I have had no problems scaling up or scaling horizontally. I have had some very large Red Hat Enterprise Linux nodes with 254 gigs of memory and a big chunky Oracle database system. We have had no problems with them. We have not had any problems with running with multiple memory cluster nodes. We have had 100 gigs network, and we had no problems. We had a high-end SAN and a high-end network, and we had no issues.
They have good integrations, and they have not had too many problems with external SAN providers. They have been fairly consistent with keeping up with everybody else and keeping their drivers good.
How are customer service and support?
They are probably one of the better ones in the industry. I can get a real answer, and I do not feel like people are breathing down my neck and saying, "I am going to close your ticket. I have not heard from you in 15 minutes." It has been a very positive experience. They have always helped us out when we have completely gone sideways.
They are very patient with the level of experience that a lot of people have. We have a significant number of junior admins who put in tickets that probably should not have been put in. They have been very patient. Overall, it has been a good and positive experience.
How would you rate customer service and support?
Which solution did I use previously and why did I switch?
Before Red Hat Enterprise Linux, I was strictly using Solaris and AIX. I never used Ubuntu. It was just straight, big-frame Unix before I went to Linux. I did not change too many platforms.
How was the initial setup?
We use Ansible to deploy Red Hat Enterprise Linux machines on VMware. That is 80% to 90% of our workload. For everything else, I have done PXE boot and kickstarts.
We are using a hybrid cloud. Our cloud providers are Azure and AWS. We work with both. The deployment on Azure and AWS was simple. We built Elasticsearch inside of Azure. It was a click-button deployment. We use TerraForm to deploy most of it, and then we have Ansible to do the rest.
I wanted to try to do more infrastructure as code, but it is hard to get traditional admins into that mindset, so it is always a mix. I deploy these servers for them with TerraForm, and then I pretend I never did, and they can do whatever with them. It then goes back into traditional life cycle management. Sometimes they delete them, and sometimes they forget about them. Satellite has helped us keep track of where everything is. It has helped us track our life cycles. It has been helpful for us.
What about the implementation team?
We have used Red Hat consultants multiple times. They helped us set a few things up and clean up our pipelines. We have been very happy with our Red Hat consultants and our last deployment of OpenShift AAP. We loved their consultants. They were fantastic.
What was our ROI?
The biggest ROI that we have seen by using Red Hat Enterprise Linux is accessibility to information for frontline support people, midline support people, and developers. There is a ton of information, and there is a ton of community support.
For us, Red Hat Enterprise Linux is a consistent platform because if we are on a customer's Rocky machine, we already know Red Hat Enterprise Linux. We can deal with that. It is a skill set that is very broad across multiple platforms. That means we can apply what we have learned and what we have been trained in. While working with the Red Hat Enterprise Linux team, we have learned best practices, and we can apply those across the board. That partnership has helped us better our internal practices whether it is Red Hat Enterprise Linux or not. That is a positive. Satellite has also been a real positive for us because we can now manage all of our systems from a single pane of glass. That is what my frontline people have been asking for. They wanted one place to patch the systems, and now they can.
What's my experience with pricing, setup cost, and licensing?
Our experience was incredibly positive because we started working with OpenShift before we were fully licensed. They knew we were going in that direction. The same thing happened with Red Hat Enterprise Linux. They knew we would buy tons of Red Hat Enterprise Linux, so they were a little bit more relaxed. We wanted a thousand licenses, and we could pick those up. We true up. Our license experience has been positive with the exception of having to deal with all of the broken-up accounts, which is as much our fault as anybody's.
My biggest complaint is that we have eight or ten different contracts. It is hard to keep track of what is on what and where we are getting the most value-add out of our benefits.
They are helping us solve that problem. We have reached out to our account executives. They will help us solve that problem. That is a huge step because that has been a problem for 15 years. It will help us consolidate and understand what we are spending across the board instead of seeing what we are spending in chunks.
OpenShift has come close to paying for itself in the first year and a half. That is an easy business case to make if you have the direct ability to show cost savings. We are getting cost savings, and we have the ability to show those cost savings. These are the two major benefits we have seen with AAP and Red Hat Enterprise Linux bits. That has been a positive for us. Red Hat Enterprise Linux AI and some of the other things they are starting to do are probably going to enable a lot of our developers to start taking advantage of them. Red Hat Enterprise Linux AI changes the belief that AI is out of reach for a normal developer.
Which other solutions did I evaluate?
We considered the idea of building this entire platform on Rocky as a free solution. It just was not cost-effective. There are hidden costs of patching and maintaining. They require care and feeding. We wanted cattle, not pets. We had a bunch of pets. Red Hat Enterprise Linux enabled us to get into that cattle methodology and mindset. Our mesh nodes are built on Red Hat Enterprise Linux 9. If my mesh node goes sideways, I do not care. I just delete the VM, redeploy it, and run my playbook. In 15 minutes, I am back up and running again. Why would I troubleshoot it? It takes time. I do not care about troubleshooting. It enables us to rinse and repeat a lot of our processes.
What other advice do I have?
People turn off too many of the tools way too often. We have a lot of room for improvement as an organization to embrace SELinux. We are still working on that. That has a significant amount of value. We want to embrace the GPG sign code in AAP. I do not want anything but approved containers and code running on our platform and our customer's platform. They have enabled us to be incredibly secure, and we are yet to fully take advantage of those offerings. It is a goal, and we are going to get there.
To a colleague who is looking at open-source, cloud-based operating systems for Linux instead of Red Hat Enterprise Linux, I would say that Red Hat Enterprise Linux-based variants are the best in my opinion. If I have a choice, I will always go for CentOS, Fedora, Rocky, or something else that is Red Hat Enterprise Linux-based. If they were not going to go with Red Hat, I would probably tell them to go with CentOS but stay behind a little bit because they do not want to be at the bleeding edge of CentOS. That relationship kind of changed when they took it to AppStream instead of a more supportive platform.
I would rate Red Hat Enterprise Linux a nine out of ten. They keep doing well, and they keep getting better. As long as they stay on the same path, I do not see us not using Red Hat Enterprise Linux in the future. It has been consistent. Why would we change?
Disclosure: My company does not have a business relationship with this vendor other than being a customer.