What is our primary use case?
It started mostly with websites and open source environments overall for development. Now, we are moving into business applications as we are migrating our ERP, which is a cp -r tree, to Linux. We are also migrating the database of SAP to SAP HANA on Red Hat Enterprise Linux.
We use RHEL versions 7 and 8. There is a bit of version 6 still lying around, but we are working on eradicating that. It is mostly RHEL Standard subscriptions, but there are a few Premium subscriptions, depending on how critical the applications are.
How has it helped my organization?
It has fulfilled all the promises or goals of different projects, not just because our internal team is strong, but also because our external partner is strong.
What is most valuable?
Satellite is an optional system which provides for extensive deployment and patch management. That is quite valuable.
We use Red Hat Enterprise Linux's tracing and monitoring tools. You don't leave them on all the time, as far as tracing is concerned. When you are sick and go to the doctor, that is when you use it, e.g., when an application is sick or things are really unexplainable. It gives you a good wealth of information. In regards to monitoring, we are using them to a point. We are using Insights and Insight Sender as well as the Performance Co-Pilot (PCP), which is more something we look at once in a while.
Other Red Hat products integrate with Red Hat Enterprise Linux very well. In fact, they integrate with pretty much everything around the universe. We are doing API calls without even knowing what an API is, i.e., towards VMware vCenter as well as Centreon. There are certain individuals who use it for free without subscription and support for Ansible in our Telco group with great success.
What needs improvement?
Linux overall needs improvement. They cannot go much beyond what Linus Torvalds's kernel implementation can do. I come from AIX, and there were very cool things in AIX that I am missing dearly, e.g., being able to support not only adding, but also reducing memory and number of processors. That is not supported on Linux right now, and it is the same for the mainstream file systems supported by Red Hat. There is no way of reducing a file system or logical volume. Whereas, in AIX, it was a shoo-in. These are the little things where we can say, "Ah, we are missing AIX for that."
We are not loving our servers anymore. If we need them, we create them. When we don't need them, we delete them. That is what they are. They are just commodities. They are just a transient product.
For how long have I used the solution?
I have been using it for nine years, since 2012.
What do I think about the stability of the solution?
The stability has been very good. However, there is a learning curve. We were running huge in-memory databases, about 2.5 terabytes of RAM, which is SAP HANA. Then, we were getting really weird problems, so we asked the app guys 20,000 times to open a ticket because we were seeing all kinds of weird timeouts and things like that on the OS side. We were saying, "It's the app. It takes forever." Finally, they said, "Oh yeah, we use a back-level thing that is buggy and creates a problem." It took us six months and four people to get that from the app guys. We were ready to kill them. That was not good. Whatever you put on Linux, make sure that you have somebody supporting it who is not dumb, or on any platform for that matter.
What do I think about the scalability of the solution?
The scalability is six terabytes. That is what we're doing. We are printing HANA servers on that scale, which are more in the 2.5 terabyte range. However, we had to create one for the migration initiative on the VMware, which was six terabytes with 112 cores. It worked, and that was it. It also works with bare-metal, but you have to be aware there are challenges in regards to drivers and things.
How are customer service and technical support?
RHEL provides features that help our speed deployment. For example, for SAP HANA, they have full-fledged support for failover clustering using Red Hat HA, which is a solution to create a vintage approach of failover clustering. They do provide extensive support for value-adds for ERP solutions.
They also provide value left, right, and center. Whenever we have a problem, they are always there. We have used both their professional services as well as their Technical Account Manager (TAM) services, which is a premium service to manage the different challenges that we have had within our business. They have always come through for us, and it is a great organization overall.
Their support is wonderful. They will go beyond what is supposed to be supported. For example, we had a ransomware attack. They went 20 times above what we were expecting of them, using software provided by them on a pro bono basis, meaning take it and do whatever you want with it, but it was not ours. That was a nice surprise. So, whenever we have needed them, they did not come with a bill. They came with support, listening, and solutions. That is what we expect of a partner, and that is what they are: a partner.
Which solution did I use previously and why did I switch?
I, for one, was managing AIX, which is a legacy Unix, as my core competency. I still do because we haven't completed the migration.
RHEL is a value-add right now. As we are migrating more payloads to containers, we are putting less Linux forethought into these container-hosting servers. You just shove your containers on top of them with your orchestrations. This may reduce our need to manage RHEL like a bunch of containers. That changes the business.
We were paying for premium SUSE support for an initial pilot of SAP HANA on the IBM POWER platform. We were stuck between an IBM organization telling us, "Go to SUSE for your support," and the SUSE organization saying, "Go to IBM for your support." So, we told them both to go away.
We are so glad that we haven't mixed the Red Hat and IBM more, because SUSE and IBM don't mix, and we were mixing them. That was prior to the merger with Red Hat. In regards to IBM's ownership of Red Hat, we are a bit wary, but we think that IBM will have the wisdom not to mess it up, but we will see. There is a risk.
How was the initial setup?
The initial setup is as straightforward as it can get for anyone who knows what they are talking about. It does require technical knowledge, because that's what it is: a technical solution. It is not something that I would give to my mother. Contrary to other people's perception of, "My mom had a problem with her Windows. Oh, put her on Linux." Yeah, no thanks. Give her a tablet, please. Tablets are pretty cool for non-techies, and even for techies to a certain extent.
For the migration from AIX, Ansible has been our savior. You do need somebody who knows Ansible, then it is more about printing your servers. So, you press on the print button, then you give it to the apps guys, but you do have to know what you are trying to aim for so the guy who is creating the Ansible Playbook codes exactly what is required with the right variables. After that, it is just a question of shoving that into production. It is pretty wonderful.
What was our ROI?
We do get a return on investment with this solution in regards to a comparative cost of ownership of going with the niche solution of IBM AIX systems and hardware. There is a tremendous difference in cost. It is about tenfold.
The integrated solution approach reduces our TCO tremendously because we are able to focus on innovation instead of operations.
What's my experience with pricing, setup cost, and licensing?
RHEL is a great place to go. They have a great thing that is not very well-known, which is called the Learning Subscription, which is a one-year all-you-can-drink access to all of their online self-paced courses as well as their certifications. While it is a premium to have the certifications as well, it is very cool to have that because you end up as a Red Hat certified engineer in a hurry. It is good to have the training because then you are fully versed in doing the Red Hat approach to the equation, which is a no-nonsense approach.
Because it is a subscription, you can go elastic. This means you can buy a year, then you can skip a year. It is not like when you buy something. You don't buy it. You are paying for the support on something, and if you don't pay for the support on something, there is no shame because there are no upfront costs. It changes the equation. However, we have such growth right now on the Linux platform that we are reusing and scavenging these licenses. From a business standpoint, not having to buy, but just having to pay for maintenance, changes a lot of the calculations.
Which other solutions did I evaluate?
We tried SUSE on the IBM POWER platform, and it was a very lonely place to be in. That was for SAP HANA migration. We are glad that we decided to be mainstream with leveraging what we already had at Red Hat Linux (over a few dead bodies now). We also leveraged the Intel x86 platform, which is very mainstream.
We are not using the Red Hat Virtualization product. We are using VMware just so we can conform to the corporate portfolio.
Our RHEL alerting and operation dashboard is not our route one right now. We have been using Centreon, which is derived from the Nagios approach, for about seven years.
With AIX, we couldn't get a single software open source to run. It was like a write-off, except for reducing a file system or logical volume in Linux.
What other advice do I have?
We are a bunch of techies here. RHEL is not managed by end users. We don't really mind the GUIs, because the first thing that we do is stop using them. We are using Ansible, which is now part of RHEL, and that can automate the living heck out of everything. For now, we are not using the Power approach, but we may in the future. We are doing a business case for that, as it would be an easy sell for some communities and the use cases are not techie-to-techies.
There is a cloud, but we have very little infrastructure as a service in the cloud right now.
It delivers to the targeted audiences. Could Red Hat Enterprise Linux be used in all types of other scenarios? Most likely. They have an embedded version for microcontrollers, i.e., things that you put into your jewelry or light switches. However, this is not what they're aiming for.
I would rate RHEL as a nine and a half (out of 10), but I will round that up to 10.
Which deployment model are you using for this solution?
Which version of this solution are you currently using?
7 and 8