I cut my teeth in Enterprise Storage for 3 decades. On and off, I get the opportunity to work on Cloud Storage as well, mostly more structured storage infrastructure services such as blocks and files, in cloud offerings on AWS, Azure and Alibaba Cloud. I am familiar with S3 operations (mostly the CRUD operations and HTTP headers stuff) too, although I have yet to go deep with S3 with Restful API. And I really wanted to work on stuff with the S3 Select when the opportunity arises. (Note: Homelab project to-do list)
Along with the experience is the enterprise mindset of designing and crafting storage infrastructure and data management practices that evolve around data. Understanding the characteristics of data and the behaviours data in motion is part of my skills repertoire, and I continue to have conversations with organizations, small and large alike every day of the week.
This week’s blog was triggered by an article by Tech Republic® Jack Wallen‘s interview with Fedora project leader Matthew Miller. I have been craning my neck waiting for the full release of Fedora 36 (which now has been pushed to May 10th 2022), and the Tech Republic®’s article, “The future of Linux: Fedora project leader weighs in” touched me. Let me set the context of my expanded commentaries here.
History of my open source experience- bringing Enterprise to the individual
I have been working with open source software for a long time. My first Linux experience was Soft Landing Linux in the early 90s. It was a bunch of diskettes I purchased online while dabbling with FreeBSD® on the sides. Even though my day job was on the SunOS, and later Solaris®, having the opportunity to build stuff and learn the enterprise ways with Sun Microsystems® hardware and software were difficult at my homelab. I did bring home a SPARCstation® 2 once but the CRT monitor almost broke my computer table at that time.
Having open source software on 386i (before x86) architecture was great (no matter how buggy they were) because I got to learn hardcore enterprise technology at home. I am a command line person, so the desktop experience does not bother me much because my OS foundation is there. Open source gave me a world I could master my skills as an individual. For an individual like me, my mindset is always on the Enterprise.
The Tech Republic interview and my reflections
I know the journey open source OSes has taken at the server (aka Enterprise) level. They are great, and are getting better and better. But at the desktop (aka consumer) level, the Linux desktop experience has been an arduous one even though the open source Linux desktop experience is so much better now. This interview reflected on that.
There were a few significant points that were brought up. Those poignant moments explained about the free software in open source projects, how consumers glazed over (if I get what Matt Miller meant) the cosmetics of the open source software without the deeper meaningful objectives of the software had me feeling empty. Many assumed that just because the software is open source, it should be free or of low costs and continue to apply a consumer mindset to the delivery and the capability of the software.
Case in point is the way I have been seeing many TrueNAS®/FreeNAS™ individuals who downloaded the free software and using them in consumer ways. That is perfectly fine but when they want to migrate their consumer experience with the TrueNAS® software to their critical business operations, things suddenly do not look so rosy anymore. From my experience, having built enterprise-grade storage solutions with open source software like ZFS on OpenSolaris/OpenIndiana, FreeNAS™ and TrueNAS® for over a decade plus gaining plenty of experience on many proprietary and software-defined storage platforms along this 30 year career, the consumer mindsets do not work well in enterprise missions.
And over the years, I have been seeing this newer generation of infrastructure people taking less and less interest in learning the enterprise ways or going deep dive into the workings of the open source platforms I have mentioned. Yet, they have lofty enterprise expectations while carrying a consumer mindset. More and more, I am seeing a greying crew of storage practitioners with enterprise experiences dealing with a new generation of organizations and end users with consumer practices and mindsets.
TrueNAS® (and other open source software) are meant for great things
TrueNAS® today has 3 different editions. CORE, Enterprise and SCALE. Each was developed to serve different markets, and therefore, should be managed as such. Along with the capabilities of these editions, one must also be ready to practice the enterprise ways when used in an enterprise setting, like using the right tool for the right expectations and outcomes.
So I was perplexed when I read user comments in the Facebook communities and the forums where users brought a consumer mindset to deploy a TrueNAS® storage in an enterprise environment, and then make a fuss (or a ruckus) of the issues caused. I can share a few examples:
- One user in a hedge fund company lost database data on a RAID-0 volume (dataset in a zpool) because he found out from a friend that RAID-0 gives the highest IOPS performance from all the RAID levels available on TrueNAS®. He went on to say that TrueNAS® is not reliable.
- One user from a cybersecurity risk management firm wanted TrueNAS® CORE to provide similar features like VMware® Tanzu where TrueNAS® is a hypervisor for both VMs and containers. And he wants to have TrueNAS® with xcp-ng hypervisor in the self-assembled x86 hardware. He is still scratching his head why TrueNAS® cannot do what he wants to do.
- Too many have asked for the TrueNAS® Enterprise license (which they were willing to pay) to run their own procured x86 hardware while iXsystems™ exercises a storage appliance business offering. I have gotten a few sarcastic remarks saying that iXsystems™ is behind time for not giving them the opportunity to expand the market for TrueNAS®. There is always the free version in TrueNAS® CORE.
- One user asked for > sustained 300K IOPS for his MySQL databases. An all-flash storage appliance was proposed and in the end, I found out that he purchased a QNAP® instead.
- One cloud hosting company complained that their NFS operations “hung” for 20+ minutes every day their Synology® NAS did a snapshot. They are now a happy iXsystems™ TrueNAS® Enterprise customer.
- More and more users are using TrueNAS® SCALE for home use purpose. The SCALE software is great and it is meant for greater things for the enterprise.
If one takes the enterprise high road, and takes the time to understand the open source approaches and business models used by iXsystems™ and TrueNAS® or other open source projects such as MinIO, Red Hat®, and more, one will realized that these projects are developed for the enterprise. Therefore, the expectations and the best practices must be of the enterprise mindset, not the consumer kind of mindset. Having installed, deployed and used these open source software as an individual with an enterprise approach and mindset, these software are developed for greater enterprise operations.
I may entirely wrong here. But collectively, I have gathered that most of the issues at the enterprise operations level could be avoided if the end users and the organizations took the effort to research the open source storage technology. Yes, some may be ignorant, and so may be blinded by other factors. But we must always remember that we can’t invest in a truck to race in a Grand Prix and we can’t purchase a Bugatti to haul your cargo van.
There are open source software that for consumers and for the enterprise. And so much the objectives and the mindsets wherever the open source solutions are deployed, operated and managed.
This blog is about having the right mindset, not just the technology. So be one with the right one.