Comment by jabl | Hacker News Reader

Comment by jabl | Hacker News Reader

Parent Full thread

For a somewhat simplistic explanation, RHEL became very successful as a tool to replace proprietary RISC Unix servers (often running Oracle DB) with x86(-64) and Linux, offering a traditional "enterprise" support model those proprietary UNIX customers were used to. So what has changed?

1. The above market is pretty much "saturated" in the sense there aren't many proprietary UNIX systems left to convert to x86+RHEL. So not much growth potential there.

2. Oracle has their own RHEL rebuild that they aggressively(?) market to Oracle customers. And I guess it makes sense if you have a server to run Oracle DB on, to also use the Oracle RHEL rebuild to avoid the situation where you have a problem and Oracle and Redhat point fingers at each other and nobody takes charge of solving the issue.

3. Cloud. Cloud is obviously big these days, and it seems for one or another reason cloud deployments often choose some other distro than RHEL, be it one of the free rebuilds (CentOS, Alma, Rocky, OEL), or something else entirely like Ubuntu.

While I'm not privy to any RH internal info, it wouldn't surprise me if they're worried about the above trends continuing. Whether this recent change to make life for rebuilds a bit harder will materially affect anything, and if so how, remains to be seen.

re 2) - we once fell into this marketing trap of "Oracle runs best on Oracle". If it only were true!

We had Oracle databases running on Oracle Solaris running on Oracle SPARC servers, because the bosses were convinced of the synergies. When the problems started (with Solaris zone clusters) it was a happy merry-go-round of fingerpointing between Oracle support groups.

Oracle is on a rather short list of companies I'll try to avoid doing business with if at all possible.

This strategy is doubling down on the real reason that 3 is killing them: price. Their pricing structure in fundamentally incompatible with cloud native workloads. We treat workloads like cattle scaling them quickly and on demand to OS counts never dreamed of decades ago.

...And when you graze your Cattle on someone else's fields, you pay through the nose for it!

Best damn business innovation ever, renting people shovels to dig themselves a hole with!

It's a classic IBM strategy that goes all the way back to the old-school days of mainframes and timesharing. And they would love to bring those days back. But they can't because AWS beat them to it and now they have to live in that world.

re:3, In a previous thread about this, someone said that RHEL only would make deals with cloud providers to run RHEL licenses prepackaged if the whole stack was RHEL, and obviously, if the cloud provider is not already running RHEL everywhere, that doesn't work out so well.

(I have no personal experience trying to run anything like this, this is purely citing a previous comment made.)

I'll (mostly) copy my reply from there: RHEL guests are supported on RHEL host but also on Windows, ESX, Amazon and Google clouds, and a few more.

Non-RHEL hosts are only supported is Red Hat has someone to talk to at the vendor or cloud provider, so that they can debug stuff that breaks only in their environment and can't be reproduced. This is done to cannot guarantee the level of support that customers expect but very few cloud provider can provide that. I am not even sure that IBM's cloud made the short list.

On other unsupported hosts we will ask the customer to reproduce on a supported host if we think the host is the culprit.

>I am not even sure that IBM's cloud made the short list.

Probably not https://www.redhat.com/en/partners/certified-cloud-and-servi...

This move doesn't fundamentally fix any of those issues.

Reply on news.ycombinator.com