What is our primary use case?
Our primary use case is to develop the servers and production. It's pretty standard usage. We have some Docker running but I haven't been involved in those environments very often. It's a standard server on minimal load and we're not using a full load with our GUI interface.
We have multiple applications running on both Windows and RHEL. The database systems are mostly MySQL. There's some Oracle but most of it is MySQL. Dealing with Red Hat is pretty straightforward. I haven't run into issues with it.
When we were running multiple versions of Java, if patches came out for both versions, we would apply the patches for both versions and usually, that could be downloaded. It was pretty simple to update those. Those are systems-supported patches. With the specific application patches, it's a little different. Normally the application administrators take care of those themselves.
If Red Hat's system is set up right, it improves the speed and performance reliability of our hardware because it doesn't use as many resources as a Windows system.
What is most valuable?
I like the fact that most of the system configuration is Namespace so it's easy to get to and easy to configure, and most of it still uses text documents. Not all of it's a menu-driven-type entry. I also like the fact that it's a very standard file system layout so it's easy to navigate.
In some instances, it provides features that help speed development. In other instances, it's standard amongst most Linux groups. You can download the main features. The file system is always a big difference. You go from a Debian-based system to Red Hat, so the file system layout is a little bit different. User-based files are located versus system-based files. RHEL keeps everything in one area and segregates it. It makes it easier to go between different organizations and still have the same policies and structure. I do like the new package manager.
It's all text-based, all command line, whereas the minimal load does not have a GUI on it. If you're used to using Windows core servers, it wouldn't be that big of a deal, but going from a Windows GUI-based system to an RHEL command-line-based system is a learning curve for most Windows administrators. A lot of strictly Windows administrators don't even want to look at a command line from Red Hat because the commands are so different from what they're used to. There is a learning curve between the two major platforms.
The application and user experience are usually pretty consistent, but that really depends on the application developer. If they're developing an application, it'll be consistent costs on infrastructure. That's not an issue between the different platforms. User experience is based on how the application developer built it. They're not all in-house and so they developed across a consistent platform. They keep everything pretty simple from the user perspective.
It enables us to deploy current applications and merging workloads across physical hardware and VMs. Virtualization and physical hardware stay consistent. Going to the cloud depends on the platform we use but it'll mostly be consistent as well. The RHEL distribution has been implemented pretty well amongst most of our cloud providers. It's pretty standard now, whether we go to Rackspace, Amazon, Azure, or even Microsoft supporting RHEL distribution. You can go to a Microsoft Azure cloud and have a Red Hat Enterprise Linux system running there. The user would probably never notice it.
For Red Hat, the bare metal and virtualized environments are pretty reliable. The only thing I don't like about Red Hat is that every time you do an update, there are patches every month and you have to reboot the system. Fortunately, it's a single reboot versus Microsoft, which likes multiple reboots, but still, you have to reboot the system. You have to reboot the server. The newer updates have kernel patches involved in them. To implement that new kernel, you have to reboot the system, and Red Hat's best policy and best practices are to reboot the system after patching.
I used the AppStream feature a couple of times. Not a whole lot because a lot of our environment is specific to what we deploy. Normally I would just deploy the bare system, adding features requested by the application administrator, and they'll download the rest of the things that they need.
We have used the tracing and monitoring tools in certain instances but not consistently. We use them for troubleshooting but not every day. We use other third-party software to monitor the system logs and alert on the issues. They will run tracing analysis of the systems. The reason we don't always use it is because of the number of servers I have to deal with and the low band power.
Automation is however you set it up. As for running a cloud-based solution, a lot of it would be automated. Going from prior experience, dealing with it before coming to this company, we did a lot of cloud deployment and it's pretty consistent and reliable and you could automate it pretty easily.
RHEL accelerated the deployment of cloud-based workloads in my previous experiences. Compared to no other solution at all, it's obviously a vast improvement. You have to worry about Windows. As soon as you bring the server up, it requires numerous patches and it'll take several reboots unless you have an image that is very patched and deployable there. Even then, every month you get new patches. Red Hat patches a lot faster than Windows and requires a single reboot. The speed of deployment is a hazard. It's almost twice as fast deploying an RHEL solution as it is a Windows solution.
What needs improvement?
I'd like to see more of NCurses type menu systems in some instances. We're dealing with SUSE Enterprise Linux, they have an NCurses menu system. It's a menu system. Even some of the higher-end Unix systems like AIX have some inner menu system where all the configuration tools are right there so your administrator doesn't have to jump through multiple directories to configure files if needed. I like the simplicity of Red Hat because it's pretty easy but having an NCurses menu when you have to get something done quickly would be nice.
For how long have I used the solution?
I started using Red Hat back in 1996. I've been using it for at least 20 years, off and on. I was hired as a Linux administrator for RHEL 6, 7, and 8, and then I changed my job positions. I'm not actively using RHEL right now.
Unfortunately, we're moving away from RHEL to Oracle Enterprise Linux in the next couple of months.
What do I think about the stability of the solution?
RHEL is a very stable product. It's been around for a long time now. It's been stable since they brought it out as an enterprise environment. It's usually not the bleeding edge of Linux. That just means it has more stability in the packaging and the repositories. They keep the bleeding edge updates and things out of it most of the time, which means if you have new features that you want to implement, you have to do some finagling to get those features in place. But it does mean the system's more stable, for the most part.
What do I think about the scalability of the solution?
It's very scalable. I haven't seen any issues with the scalability of Red Hat. I've used it in environments where we have a few hundred people to a couple of thousand people. I've never seen any issues with scalability. It's one of the big sell points of RHEL. It's as scalable as Unix.
There are around 500 developers who use it. Web-facing interfaces, it's in the thousands.
If you're using a small environment with no more than around 100 to 200 servers, one or two people can handle most of the RHEL stuff pretty quickly.
If it's a larger environment, then you're looking at staff upwards from six to 15, depending on the environment the product's being used for.
There is a system administrator to perform the initial deployment of a server to the maintenance of the server. Then there are the application developers who develop the application, write the applications, and just manage applications. In our environment, we currently have sysadmins who manage the systems. My job is to manage the actual operating system itself. Then, you have application developers, who develop applications for user-facing systems. The application managers manage those applications, not only the developed applications.
It's being used pretty extensively. It's 1,100 to 1,200 servers on one site.
We're using at least 85-90% of the features of RHEL but we don't really use Ansible that extensively. Red Hat Satellite server we're using as a primary repository in one site. Based on RHEL, we use most of these features.
Which solution did I use previously and why did I switch?
We are switching to another solution mainly because a number of databases in use are based on that system. They want to expand that database and some other products that come with switching from RHEL to LEL. That's the main reason. As I understand it, the licensing isn't that different with a more centralized approach, so convenience is a large factor.
We switched to RHEL from AIX because of the developer and the cost. AIX is usually implemented on specific hardware. IBM owned the hardware. So the cost for running AIX is a lot more expensive than running an RHEL solution, which can be run on virtual systems as well as physical systems. And x86 servers are a lot cheaper than a power system.
Open-source was also a factor in our decision to switch to RHEL. Open-source has allowed a lot of development in areas, more ingenuity, innovation, and products than other constricted OSs. My opinion is that when you're dealing with open-source, you have people who are more likely to innovate and create new things. It's easier to develop an open-source platform than it is to use a closed source platform because then you can't get to the APIs, you can't do anything in the system if you want to change things in the system to make your product more available to people.
How was the initial setup?
The initial setup was pretty straightforward. I've even set servers up at home on a pretty regular basis. I have my own lab, so I've deployed it and it's pretty straightforward. With RHEL, the setup is nice because you get a GUI, so any Windows-based user is going to be familiar with the GUI and know what to look at. They can deploy software as needed, right there from the menu. From a TextBase, you can script it to where all you have to do is run a script and it'll deploy the server quickly. It's pretty straightforward.
Personally, I wouldn't be able to speak to the installation. Having a single point is always a benefit because then you don't have to jump around multiple points to download software and to deploy your solution. The only thing about it is with Docker, a lot of times you have to go out to the Docker site to download the newest versions.
If you're running Satellite, it's even easier because all your current patches are downloaded. The iOS is already there and a lot of time is it's a straight script that you can deploy quickly. The single-point install is a good thing.
Depending on what you're running it on and what kind of equipment you're running, it can take anywhere between 20 minutes to an hour. That depends on the equipment.
What about the implementation team?
They had Unix admins on site. They were implemented to bring in the Red Hat environment because of the similarity between Unix and Red Hat.
What's my experience with pricing, setup cost, and licensing?
If you implement Ansible, that's an additional cost. If you implement Satellite, that's an additional cost to your licensing. However, the amount of licensing if you license 100 servers is actually cheaper per server than licensing 50 or 25.
Which other solutions did I evaluate?
The first one that comes to mind as a real competitor would be SUSE. It's built-in Germany. Ubuntu is a commendable product but I don't find it as reliable or as easy to administer as I do RHEL. A lot of developers like it because it's really easy. It's more geared towards a home-user environment than it is a corporate environment. The support factor for RHEL is good. If you need to call tech support, it's there.
What other advice do I have?
I have used Satellite and Ansible in other environments. Satellite integrates very well. It's built by Red Hat, so it integrates thoroughly and it allows a single point of download for all patches and any software deployments you have. You can automate server builds, if you do it right, and make things a lot easier.
Ansible can tie into Satellite and RHEL fairly easily. It allows you to build multiple types of deployments for multiple solutions, and allows a playbook-type deal. You develop a playbook and send it out and it builds a server for the user. Done.
It would speed up deployment and make it easier to manage. If you had a developer who needed to throw up a box real quick to check something, he could run a playbook, throw up a server and rather quickly do what he needed to do. Then dismiss the server and all resource reviews return back to the YUM. If it was hardware, it would be a little bit different, but if we run a virtualization environment, they return all resources back to the host. So it made matching servers and deployment a lot simpler and less work on the operations environment.
The best advice I could give is if you're going from a Windows environment to an RHEL environment, there's a learning curve that is going to be a factor during implementation management and basic administration. Your company would probably need to hire new people just to support an RHEL environment. Between SUSE and RHEL, the number of people who know SUSE very well in the US is not as high as it is in Europe. RHEL has become more of a global OS than SUSE, though they're both comparable. I would advise looking at what you need it to do and then make sure you have the infrastructure, people, and manpower to support it.
There's a huge number of resources out there. You have sites geared specifically for RHEL administration. I believe IT Central Station has some resources on its site as well. There are Usenet groups and different forums. TechRepublic has a large number of resources as well. There are numerous resources out there to ease the learning curve.
There are a lot of things I've learned over the years using RHEL. Running it as a virtual design environment where you can run multiple servers on a single hardware piece makes it a lot more cost-effective and you don't have the resource depletion as you would have with Windows. Unfortunately, Windows is a resource hog. RHEL can be set up to run very minimally, with virtually no overhead other than the applications you're using to service users.
I would rate it a nine out of ten.
Which deployment model are you using for this solution?
Which version of this solution are you currently using?
7, 7.1, 7.2 and 8.1