What is our primary use case?
I have been a sysadmin handling everything about Red Hat Enterprise Linux, primarily for on-premises environments. I have built projects based on Red Hat Enterprise Linux, including development and support projects in the companies. I have handled pretty much everything. Anything you can imagine, I have probably done with it.
How has it helped my organization?
In terms of security, it does a lot of things that most people still turn off. SELinux is turned on by default. They have pretty good firewall rules in their defaults. The audit rules always take tweaking, but, overall, it comes out of the box not too bad. I used to write scripts to harden them from there.
There are multiple ways to provision and patch. You have everything from local repositories to doing it by hand.
Their knowledge base is incredible. There is so much information out there. It has never taken me longer than 30 minutes to find an answer to anything, even very tough ones.
One company I worked for was a security company, and we did a lot of patching on everything. It was designed around security and email hosting, and uptime was pretty much whatever we wanted it to be. I have had a couple of times when the uptime was bad, but it was caused by a third-party solution. In fact, the Norton antivirus was definitely the worst. Red Hat had nothing to do with it.
What is most valuable?
I enjoy the patching processes and the way Red Hat Enterprise Linux has elements set up. I have never had a patch session fail, even when installing a thousand packages at a time.
Their implementation of Yum is effective. I have used it quite a bit to pull additional information out because, for a while, I was doing security work. It is nice to have all the security information they provide on the inside. Their security library is well-maintained. I have used it exclusively for 15 years now, and I have been nothing but happy with it.
What needs improvement?
I wish IBM would give them more leeway. IBM seems to have restricted Red Hat Enterprise Linux more since the acquisition.
The organization moved away from Red Hat because IBM introduced paywalls and additional barriers that did not exist before, which made everything a lot harder. They moved from there to the Rocky version, which is a fork of Red Hat. It is run by people who have left IBM or Red Hat engineers who left IBM. Giving Red Hat Enterprise Linux more independence could be beneficial.
For how long have I used the solution?
I have been using Red Hat Enterprise Linux exclusively for about 15 years.
What do I think about the stability of the solution?
It has demonstrated great stability, with systems running for multiple years without issues. I have no problem with uptime. It is as long as you want it to be.
What do I think about the scalability of the solution?
It offers excellent scalability. The HPC system that the organization runs is a small one, but it has 8,000 computers. Each computer has at least 24 to 72 CPUs in it, and everything runs on Red Hat Enterprise Linux or Rocky. It is eminently capable. They run jobs because they do some of the hurricane forecasting and things. The things that they run on it take 900 nodes and 70,000 CPUs. You walk into that data center, the air comes out of the floor at 50 degrees. By the time it gets through the computers, it is about 110, so they are working them hard. The room itself stays at about 90.
How are customer service and support?
Before the acquisition by IBM, support was incredible. I could directly engage with developers and get immediate assistance. It was great.
I have not had a lot of experience with them post-acquisition. At this point, the entire department is moving to Rocky. It is not a huge change for me, but a part of the move was the lack of support through IBM.
How would you rate customer service and support?
Which solution did I use previously and why did I switch?
The company we are contacted with has just moved from Red Hat to Rocky, which is Red Hat-based. It is a fork of Red Hat, so it is like all Red Hat derivatives. It is binary compatible. You can do anything you want with it. For the things the organization does with the HPC environment, it is a lot nicer because there are fewer restrictions. Open source works best for HPC environments. You have to recompile a lot of drivers and things to get things to work. Being able to do that is critical in that business.
How was the initial setup?
For the most part, upgrades and migrations are very straightforward. In one of the cases, it was very straightforward to install the OS, but it was a lot more problematic to find all the pieces that ran the underlying hardware and get those working right. We had to do a lot of testing between lots of different versions of both the OS and the hardware drivers before we found good combinations. From what I hear, going from Red Hat to Rocky was a lot cleaner than going from Red Hat 6 to Red Hat 7. That was a big change.
In terms of maintenance, for the most part, once you get it set, you can walk away from it.
What's my experience with pricing, setup cost, and licensing?
I do not have any insights, but I know why the prices went up. At the time, it made sense. I do not know what the pricing is like now. Previously, the pricing model was advantageous as it allowed unlimited installations for a single price, focusing on support. The recent introduction of paywalls complicates the cost landscape.
What other advice do I have?
I have used the web console, but I am more of a command-line person. I did not see a lot of use for it, but I have used it in the HPC world because you can do some things that are handy, such as pulling in entire groups of things and building them as a boot group. It is nice to have when you need it.
Overall, I would rate Red Hat Enterprise Linux a nine out of ten.
Which deployment model are you using for this solution?
On-premises
Disclosure: My company does not have a business relationship with this vendor other than being a customer.