Thursday, December 10, 2020

CentOS, better than "not my circus, not my monkeys"

 Unemployment isn't all fun and games, even in these plague-ridden times, and it's all too easy to look on the past with the proverbial rose-colored glasses.  But every so often, something crops up that lets you say "well, at least I'm not in the middle of that shitstorm."  Better still, sometimes that shitstorm steers far enough clear that you don't have to worry about deciding whether concern or schadenfreude about your previous co-workers should hold the day.  Today is such a day for me, with the news hitting about Red Hat making their move on CentOS Linux, particularly CentOS 8.

For those not in the know, Red Hat produces versions of Linux, and has for decades, including versions intended for long-term stability, which businesses generally prefer.  Naturally, Red Hat wants to make money for their trouble, and their chosen method has been a variety of licensing schemes.  My company used Red Hat for a number of years, on their least-expensive license scheme, but, as they asked for more money for fewer support options, we decided to look for other options.  Due to the licensing structure of the various bits of Linux, it's generally possible to take the bits of one version that you like, and re-use them in another version, so long as you're not taking anything the original version coded for themselves... and that's exactly what the people at CentOS did, starting in 2004, taking the enterprise version of Red Hat's product, stripping everything Red Hat branded or coded from it, and releasing the result as its own product, the trade-off being community-based support vs. not having to pay for a license.  I don't recall exactly when, but the trade-off made sense to my company at some point, and we switched our Linux-based servers to CentOS.

Naturally, this isn't the sort of thing the folks at Red Hat wanted to allow, but there wasn't much they could do about it.  That said, the CentOS folks had their own issues, since you don't have a lot of room for doing things like hiring people to work on your code when you're giving away the final product.  Between one thing and another, Red Hat finally offered to "partner" with the CentOS folks in 2014, in what looked like a win-win for all involved, for all there were CentOS users who worried that this was a prelude to a Microsoft-style "embrace, extend, extinguish" play.  However, as the years went on, this seemed like less and less of a concern, as everybody seemed to play nice with one another.

What changed to shift that balance, I couldn't say with certainty... it wouldn't terribly surprise me if it was somehow connected to IBM buying Red Hat last year, but that's pure speculation on my part.  However, CentOS is now shifting focus to become more of a "development branch" for Red Hat's enterprise products, which is fine as far as that goes.  However, they also decided to significantly hasten the end-of-life of their most current product, from 2029 to the end of 2021, which is a rude awakening for any IT department using CentOS Linux on any sort of scale.

For me, this is very definitely a "dodged a bullet" sort of thing.  If I were still with my 2019 employer, I would doubtless be scrabbling about, looking for long-term alternatives for the various systems I still had on CentOS, and working out what modifications I would have to make to those systems in support of that change.  If I were just unemployed from that company, I would feel the need to at least give them a heads-up, and likely a "the sooner you bring me back on, the better for all of us" sort of message.  However, that company ceased operations at the end of 2019, and the startup with the refugees from that company didn't get an opportunity to build anything significant before I was let go, so it's sort of like watching the tsunami wash over an island that all the residents have safely departed from, for me.  Still, there's plenty of other shops out there that this is going to be an issue for, so it's definitely noteworthy... if only in a "don't forget these organizations can't be counted on to keep a promise" way, if and when the day comes where I'm involved in deciding on systems and services my company will use again.