Oracle Linux?

eldavojohn writes "There have been rumors floating around of Oracle working on their own distribution of Linux. If this is true, it is widely believed that this enterprise edition of Linux would be in direct competition with Red Hat Enterprise Linux. What is spurring the rumors? Well, Oracle chief executive Larry Ellison said, 'I'd like to have a complete stack. We're missing an operating system. You could argue that it makes a lot of sense for us to look at distributing and supporting Linux.' I know that Oracle has been doing a lot more than databases recently, will they go the extra mile and create their own stripped down Linux kernel? If they do, will companies switch to database solutions that are running Oracle only software for the benefits of support and (hopefully) stability?"

Comments

Definitely has uses but..

[viniosity]

If this trend continues I wonder how many orgs would be willing to go along for the ride? Imagine a mail server running on Debian, your web server running on Sun Linux, your database server on Oracle Linux, your application server on Red Hat, etc.

All similar but different enough to drive an IT guy batty. Too much of a good thing?

Re:Definitely has uses but..

[truthsearch]

Good point. But an OS stripped down and tweaked to run Oracle will most likely have the least maintenance issues. Right now Oracle has to support their DB on multiple Linux distros, plus Solaris and Windows. If they have their own OS and push it as "preferred" they'll save their customers and themselves some support cost. I think sys admins will be happy to have their database servers built specifically for their task, plus supported directly by Oracle right down to the OS level. Oracle would be adding value to their databases, so I'm surprised this hasn't happened already.

Re:Definitely has uses but..

[iabervon]

I'd run Oracle on Oracle Linux instead of some other distro. (I wouldn't bother with Oracle for any database that didn't need its own server for disk bandwidth reasons; this applies to any server that runs a single service.)

The IT guy's main headache for a database server is going to be the interaction between the database and the OS. The issue is that the server is supposed to run best on a version of Red Hat with some weird extra things enabled. Red Hat doesn't entirely understand this stuff, because they don't use it for any other configurations. Oracle understands it (they wrote it), but they're not doing tech support for Red Hat. The OS is sufficiently different from a usual Linux box that the IT guy has no clue when things are breaking. When the company I was working for got one of these, it was further complicated because the hardware didn't come with anything set up, and came from a third vendor. So we got a machine from Dell, the OS from Red Hat, and the database program from Oracle, each shipped separately, and they couldn't be tested independantly.

I think it would make perfect sense for Oracle to distribute and support a Red Hat-derived Linux distribution exclusively for production servers. At least then there would be a vendor who would understand the thing.

Re:Definitely has uses but..

[51mon]

HP-UX use to ship with the kernel settings all correct for a small Oracle database (small then being 9 to 20GB IIRC). Of course being a conscientious system admin, it didn't stop me double checking them against the Oracle documentation each time in case some advice had changed.

This didn't happen by chance. But it meant that you could be reasonable certain no obscure kernel settings were incorrectly set (at least not by an oversight, didn't stop people setting the wrong settings when tuning).

At the time Oracle were talking with Hewlett Packard about a stripped down HP-UX to build "Oracle Servers" on PA-RISC. It made sense then, and it still makes sense, except HP-UX is no longer the "obvious choice" for an Oracle server.

To be honest, I think in the GNU/Linux world, it is choice of certified hardware that is probably as important, if not more so, for Oracle, than choice of distribution. Since I've been bitten by underdocumented, under tested, RAID hardware or Linux drivers for same (the effect is the same, no matter where the fault lies). If you are aiming for really high availability on an Oracle database, buying the solution as one stop from Oracle makes sense.

I doubt cost-wise it would be that competitive with DELL and Redhat, at least initially, but for some applications hardware cost is irrelevant compared to unplanned downtime.

Something like Debian, or Ubuntu, with long support periods, and completely freely redistributable base (with builtin rebranding -- "no Mozilla says you can't call it..." hazzles), is the obvious sort of base. Although presumably BSDs might be an option as well. Or Oracle might still want a big corporate backer for their distro variant.

OpenSolaris?

[AKAImBatman]

I'm a bit surprised that they're not considering OpenSolaris. Linux is nice, but Oracle has been supporting Sun Solaris for far longer. Using Solaris as their base kernel would allow them to provide a large number of enterprisey (lt;-technical term) features out of the box.

Not to say that 2.6 doesn't have bunches of enterprisey (<-technical term again) features, but Solaris is still a leader in that space.

Re:OpenSolaris?

[atbarboz]

I work at Oracle right now - and all base development happens on RHEL 3. We are in fact upgrading all developer and QA machines to RHEL 4 over the next couple of months. Solaris used to be the base development platform around 2 years back, but it's just a porting platform as of today.

Oracle Linux works better as a threat than reality

[dcavanaugh]

Larry likes to expand Oracle's reach by purchasing competitors. Consider the world of ERP, where he bought Peoplesoft and JD Edwards. It makes little sense for Oracle to build their own distro to compete head-on with Red Hat. It makes a lot more sense to threaten to build a Linux distro in the hopes of driving down RHAT shares, thus facilitating a takeover. Larry wanted to buy JBOSS, but Red Hat beat him to it. If he buys Red Hat, he gets JBOSS as well. And all of Red Hat's customers. Buying Red Hat would make Oracle the #1 Linux company overnight.

Besides, if Oracle tries to build their own distro, market it via their existing sales channels, and support it via their existing system, Oracle Linux will truly suck. The pricing will be outrageous, the sales process will be the "car dealership" model, and the support will be the offshore model that is not all that great. Oracle makes a great product, but they are their own worst enemy sometimes.

If I were Larry, I would create a great deal of hype about doing my own Linux distro, to soften up the price of Red Hat in anticpiation of a takeover.

Try a different approach.

[khasim]

Suppose Oracle supports their own, reduced, version of Linux (with any performance enhancements that they deem necessary). If they "partnered" with a hardware vendor, you'd have a single stop for your database server needs.

You'd get your BIOS updates, OS updates and database updates from a single company that could afford to do the testing so the load on your IT department would be reduced.

You could even order it in a cluster configuration.

But what good is a database server on its own? With a bit more work, you'd be able to buy a webserver box (hardware, OS, Apache, etc) pre-configured to hook into the database server they sold you.

From Oracle's point of view, this would be a great way to get even more of the market and to stop any gains from MySQL or others.

From the corporations' point of view, this would be a great way to reduce IT costs by reducing the load on your internal IT department.

If Oracle does it right, they'd even be able to offer you dial-on-demand DBA services for their products. Why pay 6 figures to hire an Oracle DBA when you can pay 5 figures for a DBA service contract with Oracle?

Do you know what Oracle stands for?

[MrJerryNormandinSir]

One
Raging
Asshole
Called
Larry
Ellison

Reboot process

[Genady]

sqlplus sysdba@myserver/tiger

update SYSV_INIT.INITLEVEL='6';
commit;

(or something like that. I'm a SysAdmin damnit, not a DBA)