Slashdot Mirror
Oracle to Offer RedHat Support?
rs232 writes to tell us ITP is reporting that Oracle's Larry Ellison recently called Red Hat's ability to honor their support contracts effectively into question. Taking that claim one step further, Ellison claims that Oracle will soon start offering support for Red Hat Linux users. From the article: "The reason for this move, which Oracle executives later declined to provide any real detail on, is that Red Hat isn't doing a good enough job of providing that support itself, Ellison said. 'Red Hat is too small and does not do a very good job of supporting them [customers],' he said."
(If this actually happens) This would be a very interesting turn of events. Oracle is widely credited as one of the reasons that Red Hat was able to break into the enterprise. If Oracle goes its own way, it will be interesting to see how Red Hat works through the challenge. On the other hand, supporting a full-fledged distribution is easier said than done.. May be Oracle is just posturing to get a better deal out of Red Hat.
Rugged Laptop Guide
Because it means that Redhat is doing a good job and they need to grow to be able to satisfy more clients.
So, does this mean Oracle will support MySQL which is part of the Red Hat distribution?
Rugged Laptop Guide
hijack someone else Ellison! Can this guy be any more of a dick?
Ok Ellison is dissatisfied with red hat support. It would have been worse if actual OS users were. Like that other operating system's users sometimes are...
---- MISSING MISCELLANEOUS DATA SEGMENT --- [sigdash] trolololol
Given the effort required to be able to offer support on a third party distro I wonder if over time Oracle will come to the conclusion they can provide their own distro as easily as carry out support for distro over which they have no/limited control.
Either that or will Oracle end up buying RH?
--- Users are like bacteria -> Each one causing a thousand tiny crises until the host finally gives up and dies.
This is really rather funny - Oracle's support is typically dreadful, so now they will further stretch their already thin support resources a little more to bring you even *less* support per dollar!
Is it possible to link to something better than a site that's 50% advertisements and a 14 sentence article? Surely there's a better site than this.
Is this the best we can do?
Thank you for your attention. You may now return to your regularly scheduled modding down.
-- Sig down
Wikipedia says Red Hat has 1,200-1,300 employees. Of those, I suspect a few hundred are going to support.
Here's the rumor I've heard: (Can't name the source, sorry.)
If a single mega-company were to migrate to Linux and rely on Red Hat for support, it would completely consume all of Red Hat's support resources, and then some. The rumor goes that this is one of the main problems with large companies that want to move to Linux: the support capacity simply isn't there.
So, the reasoning goes, Red Hat is actually glad when projects like CentOS and Oracle support take off: Red Hat knows that it can't support everybody, it knows that it needs for it's platform to "win," it knows that there is incredible value in winning alone, and so: These developments are all good for Red Hat.
After a little research, I find this article that supports what I've heard.
A lot of us are thinking about these things in terms of home users. We don't give a damn for support- we just fix it ourselves, service it ourselves. It's part of owning a computer. But in the business, I understand they think about things differently: Support becomes a primary thing. It's not optional, even when you have internal IT people on staff.
With Ellison... Red Hat is unfortunately not meeting the needs of its users. Although we agree, our reasoning is different.
A significant part of my job is Linux sysadmin work, using licensed Red Hat Enterprise products. The tools are (for the most part) useful, reliable, and complete. The problem is, the enterprise distros are severely lacking in their packages and features.
Recently, while building a distributed mail system (multiple servers in the mail chain, multidomain support and virtual mailboxes) on RHEL4,
The recommended version for mail and database servers (Enterprise Server) does indeed have packages for Postfix (our preferred mail app) and MySQL available, but none of the Postfix packages have MySQL support enabled (Postfix has good MySQL support, including DB connection caching through a proxy interface). This effectively meant that none of the dozens of excellent mail administration tools out there would be available to us, and we couldn't put together a mail system that didn't rely on flat files in some fashion or another, without setting up parallel services (LDAP) solely to support mail services.
I built the server once on Red Hat ES and when all was said and done, I ended up with seven major components having to be compiled either from source, or rebuilt RPMs with modified spec files and/or compile flags. This doesn't bother me, except for the fact that the whole reason my employer pays thousands upon thousands of dollars for an enterprise Linux was so that we could stick to standard packages, so that if a particular machine has issues, we don't have to rely on one person to know what's going on.
I can't imagine we're the only paying client Red Hat has that wants to run a mail server that relies on a database server. It wouldn't chagrin me to change mail server or database packages (I've used most of the common ones), but looking deeper just led me to the realization that no matter which packages or paths I took, I'd still be stuck with the same issues.
Until Red Hat gains better flexibility, timeliness, and awareness of their client needs, perhaps Ellison is on the ball with his visions of supporting the clients directly. I'm guessing he won't be supporting MySQL, though. And after rebuilding the server on Debian stable, with all features we desired being available in the core distro, we're happier.
And I'm the only guy here who groks Debian well enough to run it, sigh.
Rock is dead. Long live scissors and paper!
Like big is good? I don't know how many employees Oracle has, but I can say this: The number of employees in a company is not related to how good the support and/or products are.
Let me count the ways:
I'd venture to guess more than 3/4 of its technical staff is dedicated to writing useless bug-ridden java guis (each requiring differing versions of java) with absolutely no interoperability between them. None of them can be scripted and they're all pieces of shit.
And let's not start with sqlplus. You think they could just hire one guy who may be able to put some readline support in there so it could get with the times.
Another good example is security. How many employees does oracle have dedicated to their security team? I'd venture to guess they have one monkey. Not even a person. Do I need to bring up the unpatched vulnerabilities that are hundreds of days old?
Now how about bug fixing? Anyone ever upgrade a production Oracle instance? No? You know why? Because you fucking can't. You have to wait until the latest patch has at least 1 year of testing because upgrades, even minor bug fixes, break in spectacular ways. So, because noone installs them, there's never any testing.
"What Red Hat does is every time to fix a bug you have to upgrade the operating system. They dont support old versions but just bug fixes That is not proper enterprise support and I think our customers are demanding that and I think you will see that coming from us."
I'm pretty sure this guy doesn't have a clue what he's talking about. Since RedHat soesn't change software versions after a release, but instead backports security and bugfixes to the released version, what older versions is he referring to?
Ubuntu: If at first you don't succeed, blindly slap a sudo in front of it
It seems that teh two extreems would be: Oracle goes full speed ahead with this, fatally wounding RedHat, eventually causing a void i OSS land, or RedHat stands up to the challenge and emerges stronger than before. Either extreme has their pros and cons however.
"Thanks for all the money you paid to us. We've used it to buy off ISO among other things" -Microsoft
I have talked to our developers and Product Management the last days and unfortunately we currently couldn't allocate enough developer ressources to get this issue fixed.
I will try to check when the currently estimated time for this fix is. -- Response of 18 Jan to support query filed 4 Aug 2005, still no engineer assigned...
On average we get a 6 month delay before the report reaches an engineer, and when it does the first thing we get asked is if the problem is still occurring (read fixed this yourselves yet?). Don't get me wrong. I love Red Hat and the work it does. We took on RHEL V4 instead of FC for the core services of our company, primarily for the support aspect. Out of the several support requests filed we only have had prompt decent support for one of them - and that was only because their web support had gone down and they were taking phone support. It really makes me wonder what the benefit of RHEL is over FC if support is near non-existant. Or is some big corporate with RHEL rolled out across all its servers consuming all of Red Hat's support staff, denying the small fries any look in to support?
No wonder Oracle are looking to move in
Maybe it's just me, but given the inability of Oracle to fix major vulnerabilities after years of documentation of same ... maybe MicroSoft is about to assert that, since Oracle's support for security is so poor, that they will begin provided that service.
While I completely agree with Ellison, insomuchas my enterprise-level experiences with Red Hat's support have been awful, there's another side to this coin. I don't think it's been any secret in the industry that Oracle has been displeased with Red Hat, and vice-versa, however, the outstanding question has been how they would proceed in addressing these concerns. Would they buy Red Hat? Partner more with Novell? Release their own Linux distro? Or, as this would seem to indicate, something else uniquely Oracle?
The big question here is, in my opinion, what does this say about Novell and Oracle in the enterprise? It could be argued that Oracle had already invested so much time and effort into nuturing their product line on Red Hat that a move to SUSE would be cumbersome. But, still, I would argue that Oracle's better move would be to deepen the Novell relationship. Novell has shown a consistent committment to enterprise products, Oracle included - and has the track record of good enterprise support.
Personally, I can only say that I believe a move like this on Oracle's part would only serve to strengthen the position of Linux in the enterprise. As I alluded to above, the largest hurdle Red Hat could not overcome in my enterprise was poor support - something Oracle could easily address. So, in the end, it's a win for the industry...
But, why not just buy Red Hat? And, to my original question, how much does this hurt Novell?
"Adventure? Excitement? A Jedi craves not these things."
i'm a medium size company with a number of Redhat Enterprise subscriptions running in-house. We've had great service from their customer support. A large corporation like oracle should have a good number of certified employees anyways.
Redhate Enterprise support is aggressivly priced compared to other players in the enterprise (IBM(AIX), HP(HP-UX), Oracle, Microsoft, etc.). Staffing at most of these vendors can be split into sales, support, programming, r&d, and management. Redhat's income stream will dictate how fast they can grow and how many people they employ.
:) ), and Dell(who's been able to overcome this of late).
The danger is to grow faster than your organization can absorb (so you don't have former janitors as VPs of development). If you do, quality and customer satisfaction will suffer. Some great examples are Leading Edge(anyone remember these guys?), Gateway 2000(who knew signing 900 retail leases within a couple of years could kill you?
So here's where it becomes interesting. You're potentially underfunded by your licensing model and you're seeing growth in the folks buying your service. Do you cut costs(layoff), finance expansion (go in debt to grow), or raise prices? These situations are when the CEO & CFO actually earn their paycheck. I'll be interested to see how Redhat responds.
STFU & GBTW
There are at least two senses of "support" here, which are hand-holding and actual bugfix/upgrade code changing. Answering the phones is the easy one, although dismal performance by various cost-cutting, outsourcing big vendors can obscure that point.
So the real question is indeed, as already noted in this thread, will Oracle code, package, and support a particular Linux distro? I think it has to go that way. Here are two reasons why.
1. Enterprises use huge application-oriented technology stacks -- hardware, OS, DBMS, app server, OLTP apps, analytics, etc., etc. They increasingly resist paying "value prices" for all those layers. Thus, each vendor wants ITS tiers to be value-priced, while the other layers are commoditized, both to free up money for that vendor, and to generally undermine the other big companies. Sun likes giving away DBMS. SAP is pushing cheap DBMS. Microsoft introduced low-cost DBMS. And so Oracle needs to strike back by, for example, ensuring that the OS gets commoditized.
2. Oracle code is what Scott McNealy would call "a big hairball". Customers need to be protected from the complexity. Integrating the DBMS and OS is a potential way to do that.
To err is human. To forgive is good system design.
I've worked with "enterprise" software for the past 8 years. My experience is that no vendor fixes everything we consider broken, and the largest vendors fix the least for us. The best overall support we get is from a 3-5 person company supporting a custom application they wrote for us. As far as COTS software is concerned, we've been working with an "up and coming" vendor (living on VC cash) who has been pretty responsive. The two largest software companies in the world have been little help to us, in terms of providing fixes to code that we consider broken.
Someone is complaining about RedHat support? And that someone is Oracle? Puh leeze!
I have yet to be impressed with the quality and responsiveness of enterprise vendors.
But Herr Heisenberg, how does the electron know when I'm looking?
This is what healthy capitalism at work looks like. RedHat and Oracle will compete at the business of providing support to RedHat users. If Oracle turns out to be better at providing RedHat support than RedHat themselves are, then RedHat will be forced to improve their support quality and customers will benefit. And if Oracle turns out not to be better than RedHat at providing RedHat support... well, Larry Ellison will look stupid in public. As usual.
I saw the tag "fud" for this article. Sorry, but this is not fud, it is the truth. You can give those standard Linux zealot lines about how if we had given more resources to it, had more, smarter sysadmins with better experience and so on and so forth that it would work. But the managers do not want to hear that, they are running a business, they are not in the Linux evangelism business. The reason they liked the idea of a switch is Red Hat on Intel is generally cheaper than Solaris on Sun boxes, and it would allow us to standardize on one UNIX platform. But there were just too many problems.
I am a Linux zealot myself, at home I have a Debian with no non-free software, not even non-free Java. But business does not think about that. The Linux kernel core team (Torvalds, Morton etc.) seem to have the strategy of competing for the high-end market with Microsoft and Sun (and some IBM lines, although IBM stands to benefit from Linux in other of its product lines). This seems like a good strategy to me since the high-end market seems up-for-grabs nowadays. Business feeling comfortable with Oracle running on a business-friendly distribution like Red Hat is essential. There are plenty of SQL Server databases running on Windows in production in Fortune 500 companies, how many Oracle on Red Hat's are there? This is essential. The worst-case scenario is it is still not there yet, Sun collapses, and Microsoft swallows up the market.
I am not just all talk - my home desktop is Debian with no non-free software. I evangelize Linux at work. I sent checks to the Free Software Foundation. I write GPL software. But this is not fud, this is reality that must be faced, and business feeling comfortable with Oracle on Red Hat is a must. Someone commented that Oracle support sucks and will they do better than Red Hat? Well, I don't know one way or the other as our DBA is who calls Oracle all the time. But this is important for Oracle, and Red Hat, and Linux and the whole free software community to get right.
Will Oracle end up buying RH?
Oracle has a market cap of less than $80B. It has declined about 25% in the last five years.
Redhat has a market cap of just over $4B. It has increased about 700% in the last five years.
The market summaries of these two companies may indeed be on a collision course. But it is because they are going in opposite directions.
Redhat without Oracle doesn't look so bad. Oracle with Linux is a struggle. Oracle without Linux is a death sentence.
If I have a problem I can quickly find if other people have the same problem. That's a great tool.
? id=193077
But the problem is that Oracle goes through bugzilla and marks all the bugs that affect Oracle private. For example I was researching NFS bugs and at first I had access to this bug but then Oracle hid it.
https://bugzilla.redhat.com/bugzilla/show_bug.cgi
Honesty about bugs is the best thing about Open Source and RedHat. Oracle is not ever going to learn that. SuSE is just now starting to learn that. They still don't have their enterprise products set up in bugzilla.
It's taken as faith in the open source community that the "give the software away, charge for support" business model is viable, but why spend all this time developing an OS if anyone else can best your support offerings? Red Hat has wasted enourmous amounts of money on the kernel if any johnny-come-lately can best their support offerings and contribute nothing? What's the incentive to give away your code now?
E = m c^3 Don't drink and derive E = m c^3
As long as your are comparing stock prices, it is fascinating to see the synchronization of these two companies.
Or at least read the full article and no just the title. Oracle says RedHat isn't doing good enough, so that don't mean RedHat's doing good!...
Please read the full stories...ghostbar page.
What a fscking disaster that would be. As bad a MSFT integreating IE into Windows.
"I don't know, therefore Aliens" Wafflebox1
This makes a lot of sense. I could see them using a stripped down version of Linux as a DBMS application server. It wouldn't be the first time it has been done. Think of the AS/400, and how widely it is still used to this day.
A slip of the foot you may soon recover, but a slip of the tongue you may never get over. -Benjamin Franklin
Oracle runs just fine on RedHat. We have thousands of RedHat boxes running Oracle for hundreds of customers at my site alone for at least the last 3 years. Multi-tier, RAC, etc.
How much more integrated can it get, shy of running in kernel space? There's already an Oracle-specific file system which is used in RAC installs. It's the only filesystem of its kind which is in the main Linux kernel.
I'm not really sure how much deeper it's likely to get.
The real problem with Red-Hat is the legacy of its package management system; that turd's been polished so hard that you can see your face in it, but nevertheless remains a turd.
RPM is hardly legacy. It's pretty damn feature-complete. Care to explain what the issue with it is?
THIS THING CAN TURN ON A DIME, MACROSSZERO STYLE ALSO FUCK BETA, ~NYORON
If you think about, there are a lot of cases of OS-DBMS integration, or at least highly OS-aware implementations. Examples include Teradata, mainframe DB2, data warehouse appliances, the AS/400 case mentioned in another note -- and arguably Oracle itself! Unlike other portable DBMS vendors, Oracle does a lot of OS-specific integration/interface work for each platform it supports.
o n/.
I posted a little support for that argument at http://www.dbms2.com/2006/07/09/os-dbms-integrati
To err is human. To forgive is good system design.
Are we still supposed to pay attention to "Feed me money" Ellison ? I mean the guy's a self-proclaimed prick and he just loves bashing everyone else because he has money and they don't. He's like Steve Jobs meets Bill Gates meets Ross Perot meets Michael Eisner.... or something like that. If he can humiliate Red Hat enough to make their market value plummet, he can buy it up for pennies and further strengthen his grip on the mindless corporate data industry. He'd probably drop the linux moniker and call it Ellison Enterprise OS, maybe have his gloaty mug as a splash screen. Ellison OS + Oracle Database to steal some glitter away from Windows + SQL Server.. not really to accomplish anything meaningful, just so he could take a shot at Microsoft and giggle until his vain face cracks.
-Billco, Fnarg.com
RHEL is huge. Really, massively fucking huge.
It encompasses about 4 times as much stuff as you'd see in a Solaris install, and it supports a much wider range of hardware.
Red Hat should change their support contract structure.
Tier 1: Security updates and access to beta patches
Tier 2: Customer support for the following packages:
* The Kernel (hardware/stability issues)
* Bugs in RH-written tools (Anaconda, Kudzu, system-config-*, rhgb, etc.)
* Bugs in RH-modified packages relating to basic system services (firstboot, sysvinit, auditing, etc.)
Tier 3: Everything in Tier 2 + packages in "Base System", Dev Tools, Net Servers, and the Red Hat-branded packages (Postgres, JBoss, ClusterFS, etc.)
Tier 4: Everything in Tier 3 + the rest of the distro (mostly user-land stuff)
Most businesses only really need Tier 1 or 2. They probably have their own ideas about everything else they want to run (Oracle, MySQL, SAP, whatever). And anything beyond Tier 2 is really outside a lot of what the RH developers focus on, so the support is going to be shitty anyway).
By offering Tier 3 at the current offering price, Tier 4 at an increased price, and Tier 2 somewhere in between the current supported and un-supported rates, I think Red Hat could improve their responsiveness and be able to hire more people to improve support's response time.
As it is, their support is probably tied up in bugs relating to some included packages that aren't "core" but they have to support anyway by the virtue of including them. Such a tiering structure could make better use of support resources.
THIS THING CAN TURN ON A DIME, MACROSSZERO STYLE ALSO FUCK BETA, ~NYORON
I'm not a DBA by any stretch.
:-/ Excuse me for being picky about package selection!
Yet I've gotten Oracle 9i and 10g humming quite nicely on RHEL3 and 4 (WS mind you), Fedora, and SLES 8 (United Linux). On x86 and x86_64. 9iR2 on RHEL 3 was a bit of a pain in the ass on AMD64 during install time, but once you got over a few humps you didn't have to worry about it again.
Solaris installs usually went smoother, but woe unto thee who neglected to install 32-bit userland packages on a 64-bit install.
THIS THING CAN TURN ON A DIME, MACROSSZERO STYLE ALSO FUCK BETA, ~NYORON
The day Ellison gets the five RHEL3 bugs I have in bugzilla (that RedHat is currently ignoring) fixed and errata rolled out is the day I believe that he can actually pull it off.
Here's the future of commercial software on Linux: all you will have to care about is the kernel version, if you are a commercial software vendor.
Why?
Because you'll ship your own glibc with your product, and all other standard libraries, and whatever command line utilities your product uses in its start up scripts, including a bash shell. ls. chmod. I've already got a commercial product in production that is built this way, and includes those utils I mentioned. It also comes with its own perl 5.8, linked against its own libc and libyouname it.
That is the future of commercial software on Linux. Vendors will start shipping their own distro-minus-kernel with their products, because they know it will work. Watch it happen. It's already started. Oracle will do it too, eventually.
Edith Keeler Must Die
Oracle says they plan to support Redhat. Not SUSE. Not some Oracle distro. So that right there is a stamp of approval on the entire Redhat "platform" if you will. Now there will be less fear that Oracle might make another distro its favorite soon- they would not hire a bunch of people to support Redhat if they planned to move to SUSE next month. Also this gives the Redhat+Oracle platform something you can't get with Solaris+Oracle or Windows+Oracle- a one stop shop for support. Redhat will be the ONLY OS that Oracle can completely provide support for. That means as of now Redhat is the best platform for Oracle. Period.
Oracle plans to support Redhat. Not CentOS or Fedora or some other free Redhat. That means if someone wants a solution for Oracle supported Redhat they still have to BUY Redhat's OS from the company. There might be some people (in fact I know one for sure) that might be holding out on switching to a Redhat Oracle solution (from a Solaris or Windows one) because they want support from a company far bigger than Redhat (like Sun and MS are). Now they have that. Plus I would not be surprised if many companies (do to ignorance, comfort, whatever) double dip- buy both Redhat and Oracle Redhat support. This can only grow Redhat's marketshare!
Its a win-win for Redhat- there platform becomes more stable and accepted, they will maybe get more people to buy their OS (that would prefer Oracle support compared to support from an OS vender like Microsoft or Redhat) and they get tons of free press.
I would be mad if I was at Novell or Sun today.
Open Source Sushi
...when someone would comment on Red Hat support.
Has anyone had any first-hand experience with Red Hat support themselves?
I would not be surprised to hear that it is abysmal.
I know it's their bread-and-butter but given the number of packages that they have to support and given that they only actually wrote a handful of them and given that they never seemed to have a very large support staff, I wonder if this is the tip of an iceberg.
Go RTFA. It doesn't say "Oracle will support Fedora" or "Oracle will support CentOS." It says "Oracle will support Redhat." As in the Redhat OS that you have to pay to get a copy of. It seems even Ellison doesn't get it either. To quote the article:
"We can just take Red Hat's intellectual property and make it ours, they just don't have it."
Um.....no you can't Larry. Redhat owns its name. It owns its logos. It uses these to maintain its control over their OS. You can't just take Redhat and stick it in a box and sell it as your own. Redhat will sue you into the ground for using their trademarks without permission. Then suddenly Redhat's money would come from litigation- much of the market uses that as a business model.
With this move Ellison is making Redhat's name (which he does not control) more valuable. That means more money for Redhat in the long run. One again he is trying to rule the world but ends up shooting himself in the foot. OSS is fine, Oracle's leader's grasp of trademark is not....
Open Source Sushi
Larry ought to try submitting a few hundred metalink tickets before he decides to dis anyone else's support. Oracle support is the KING of the classic support shuffle:
1) User submits ticket, giving detailed information of the exact module and section of code that is causing the problem.
2) Support immediately responds with a canned message that says they are working on it.
3) 24 hours go by with no further response, so user pings ticket.
4) Support asks user to post a pile of log output, most of which can have nothing to do with the problem.
5) 24 hours go by with no response, so user pings ticket.
6) Support says they have engaged a specialist and are waiting for a reply.
7) 24 hours go by with no response, so user pings ticket.
8) Support says the original analyst is unavailable, so they are passing the ticket to a new analyst who will find out what is going on.
9) 24 hours go by with no response, so user pings ticket.
10) Support asks user to post a pile of log output, some of which has already been posted, and what is new is for modules that you aren't even running.
11) User's Project Manager wants to know WTF is going on and why the f@#$%&* system isn't running yet.
12) Project Manager complains to his VP who claims to have a tight relationship with Oracle upper mgmt.
13) VP calls his buds in Oracle Sales and asks WTF is going on.
14) Oracle Sales schedules a conference call with the Oracle VP of Who-The-F*%$-Knows for a week from next Thursday.
15) User says "screw it!" and either downloads an open source module that does the same thing (but correctly), or just writes the patch himself.
16) User's VP & Oracle VP schedule a golf outing and a night in an NBA box.
17) Deal is cut to buy another $1M worth of Oracle SW plus 22% for Support.
How much more integrated can it get, shy of running in kernel space? There's already an Oracle-specific file system which is used in RAC installs. It's the only filesystem of its kind which is in the main Linux kernel.
/home, /var, etc could be on it, just like SGI's xfs or IBM's jfs.
You mentioned it: run it kernel space. Or rely on system calls only available to Linux.
Sure, Oracle wrote OCFS* (from knowledge they gained by purchasing DEC Rdb), but that does not mean that OCFS* can only be used by Oracle. It's a filesystem "just" like any other FS, in that
"I don't know, therefore Aliens" Wafflebox1
Oracle wants to cut into RedHat's revenue stream, to punish them for offering JBoss which effectively competes with Oracle's middleware offering.
Oracle starts offering Oracle support.
:( So anything that forces them to compete and rethink their business approach is fine with me.
We spend well north of 300k per year with Oracle, and I've been disgusted with the support we receive from them. Coming from a mysql / Postgresql background, I was expecting a lot more when I started working with and supporting Oracle systems, but Oracle's support staff are consistently hard to understand and not able to function when your problem falls outside their script. Escalation can be time consuming and even then you're not guaranteed a solution.
If the answer isn't in metalink, you're in trouble.
A couple of weeks ago we ran into a problem with a RAC cluster. After 3 hours of downtime, we logged a call with Veritas as Oracle were insisting that Veritas was the problem. I really wished we logged that call a LOT earlier... The guy at Veritas took about 2 minutes to explain which Oracle component was at fault and how to fix it.
Having said that, Red Hat support is pretty appalling too. I've had some classic responses to support questions from them, including advice NOT to hotswap disks on an HP DL380 (despite it being designed for this).
Mostly, I just dislike Red Hat lately because of their draconian licensing policies on some of their products... I can't even get eval versions of products that have my code in them
Actually it wasn't DEC Rdb they got it from, it was the DLM (Distributed Lock Manager) code from Tru64 Alpha's TruCluster that some twit in Compaq sold them. Sometime round 1998 if I remember right. I worked for Compaq in the (ex-DEC) UK Unix Support Group at the time and remember being horrified when I found out. At the time Oracle relied heavily on the DLM + CFS (Cluster File System) in Tru64 Unix to be able to run multiple database instances for a single database on different systems (Oracle Parallel Server).
My worst fear was that they would take the DLM, built their own platform neutral CFS implementation, and Tru64's DLM + CFS would become irrelevant. As it turned out Compaq had plans to kill off Alpha anyway, but the rest of my prediction came true.
OCFS is now old hat, and today Oracle have ASM. Their own, integrated, built in DLM and CFS implementation that gives them complete platform neutrality. Great for Oracle, but HP (ex Compaq, ex Digital) really shot their foot off on that one.
In fact ASM pretty much eliminates the need for any Vendor cluster filesystem. What other products are there that use CFS + DLM in the way that RAC does? I can't think of any, which also makes Veritas one of the big losers in this story.
Ah, you're right, I forgot all about that.
But wasn't TruCluster a reimplimentation of the VAXcluster layered product?
My worst fear was that they would take the DLM, built their own platform neutral CFS implementation, and Tru64's DLM + CFS would become irrelevant.
When a product like VAXclusters and DLM have been around for 20+ years, you've got to figure that *someone* is eventually going to reimplement it.
I've always been surprised that IBM & Sun & MSFT never reimplemented DLM to make VAX-like clusters. MSFT certainly has never had any qualms about "borrowing" other companies' ideas. Patent protection and NIH Syndrome are the only reasons I can think of why it took this long...
In fact ASM pretty much eliminates the need for any Vendor cluster filesystem. What other products are there that use CFS + DLM in the way that RAC does? I can't think of any, which also makes Veritas one of the big losers in this story.
I, as a long-time OpenVMS user, can't imagine not having a cluster-aware FS. How do database apps running on RAC nodes share scripts, flat files, etc without a DLM and cluster-aware FS?
"I don't know, therefore Aliens" Wafflebox1
Oracle would benefit from this enormously, but they'll make a lot of enemies in the process.
You're right, Macka, and Oracle has been on both sides of that fence in the past.
...
On the one hand, the night the Sun/AOL/Netscape deal was announced, I was horrified at the inevitable train wreck ahead. I had pretty good relationships with both AOL and Oracle in those days, and I emailed Larry in the middle of the night urging him to intervene with a counteroffer. He shot back that he didn't want to annoy Sun.
On the other hand, it's not long after that that Oracle tried Raw Iron, which was to be a DBMS that didn't require an underlying OS.
What's happened in the meantime is that Sun is giving away dev tools, app servers, and even now shipping open source DBMS with an enthusiasm rivalling that with which they partner with Oracle. So Oracle really has very little left to lose in terms of loyalty from or cooperation with Sun.
Similar stories would be true for other vendors. E.g., HP is in bed with everybody.
The industry is ever more promiscuous. Figuratively speaking, at least; literal promiscuity probably peaked in the 1980s, or at least it did for me
To err is human. To forgive is good system design.
I agree, managing multiple boxes is where the money gets charged, because that is (historically) something that only enterprises wanted to do -and it is where all enterprise-grade distros ask for money. Now that anyone with Xen or VMWare player can have a heterogenous cluster of linux distros on their laptop, everyone needs distributed cluster admin.
Novell's red carpet stuff is free (well, "evaluation" free) for two systems. Since I cannot get the update client that shipped with SuSE10.1 to work properly, I haven't even looked at the zenworks server. Distros should know never to ship with broken update applications, as it is one thing you can't recover from.Have you looked at LCFG for managing many linux systems. It is one of the very-large-scale linux desktop management tools, used for very large clusters including european grid installations, edinburgh university's computing department's linux infrastructure, etc.
I just did a "quick" job.
By quick, I mean two days billable.
A client was demonstrating an application using Oracle running on EL3. Hardware platform was a SUN v40z, with 8GB of memory. The client had a "simple" problem -- the sysem was only using half of the available memory.
Solution? Of course its obvious. Simply deploy the large memory kernel. But, they had three Oracle people on site, who were not familiar. The client had brought in someone else, who had no clue. I was happy, because I get to bill at emergency rates (a demo was scheduled for less than a week away). The client also wanted me to look at kernel tuning for Oracle.
If Oracle starts providing this service, it will, of course, cut me out of the loop. But I don't think it can change right away. Oracle has to provide a lot of internal training first. I expect that there will be work "in them thar hills" for the next two years...
Ratboy
Just another "Cubible(sic) Joe" 2 17 3061
Yep. Though its not quite as good as the cluster filesystem on OpenVMS. Any cluster member can read files (of 64k or larger) direct from shared storage, but asynchronous writes are still funneled over the cluster interconnect to an AdvFS domain server. So its important for optimum performance (and to avoid flooding the Memory Channel interconnect) to align the server for a particular AdvFS domain with the cluster member doing the most I/O to that domain. However, AdvFS on TruCluster does support Direct I/O, so if an application has multiple instances running on different cluster members, and they can communicate with each other to coordinate I/O as Oracle RAC does over the cluster interconnect (when its not using ASM) then the app can open files with using O_DIRECTIO and write to them directly, bypassing the AdvFS domain server and AdvFS buffer cache.
The only other Unix Cluster Filesystem I've seen that I would compare to OpenVMS and doesn't have this asynchronous limitation is the Symetrical Cluster File System from Polyserve. Though I'm a bit wary of proposing them as a solution to Linux customers as their products also run on MS Windows, and I'd be very surprised if MS doesn't gobble them up at some point in the future and dump the Linux port.
Ah, but you, like me, have "grown up" with an OS that implements SSI (Single System Image) natively. We know the benefits of having cluster common system files and application directories. For example a common account/password database without having to resort to NFS or similar. But the rest of the industry have never experienced this and frankly don't know what they're missing. As you may be aware, HP were planning to port TruCluster functionality from Tru64 over to HP-UX, but last year it got canned. My understanding of events is that this decision was customer driven. I was told that HP held workshops with their top customers and asked them what it was they wanted. The message they got back from that was their customers didn't like the fact HP were about to change HP-UX into something almost unrecognizable that would force the majority of their customer base to re-learn what they already knew. They wanted to move forward with a "unix standard" CFS implementation (i.e. Veritas) and didn't see SSI as a big enough reason to make such a huge change. Veritas had a product almost ready that meant HP-UX could get to market with a CFS a year ahead of when the TruCluster port was due to finish, so the decision by HP to drop it was a no brainer. The majority of their customers just didn't want it! Once TruCluster finally dies (it's not quite yet, you can still buy it) I doubt you will ever see another another commercial Unix SSI cluster solution. The industry just doesn't seem to see the need.
Well, I'm no Oracle expert, but as I understand it if you're running Oracle ASM, then from inside Oracle you can change directory, and view files (cd, ls) from any node in the cluster as if you were running an OS native cluster filesystem. So once oracle is up and running it can look to itself for any node specific scripts it needs. Plus it has its own DLM.
As for app binaries, Oracle recommends you have an 18GB filesystem (disk or partition) that is private to each RAC cluster member and the Oracle binaries and startup files are located in there. The Oracle RAC installation process then copies parts of the Oracle kit that need to be cluster common into a common area which I think exists inside of ASM. 18GB is nothing these days, and you can easily create this on a SAN and make it private to a system using selective storage presentation and SAN Zoning from the SAN of your choice. So having multiple copies of some Oracle files on your cluster is not considered a problem.
You should try using a AS/400, eh, System 5, eh iSeries, eh, that thingie over there, someday.
IBM defintively got something right here, and - yes - DB2/400 is embedded in the OS.
--
Thorbjørn Ravn Andersen "...and...Tubular Bells!"
Submit TAR as Pri one when it is a showstopper - somebody will call you back within two hours. Never had problems with this.
You may get even patches for your box in a week or two. Offcourse you have to have mass for it.
It's like Microsoft is doing patches still for NT4, but they are not for public (behind NT4 support is the fact, that lots of US mil stuff is still running on it).
Loved your post ;-) I have seen it in reality many times.
I was a product manager at Oracle. Bugs were identified by support and then triaged for importance by product management (probably still works that way, but I admit that I am out of the loop now). If your bug is low impact (i.e. I not encountered by 99% of users) and has not been reported by a high priority customer (i.e. one that is about to spend big $ on new licences or where the CEO plays golf with Larry) you are going to find your bug at the back of the line. My advice - Kick up a fuss. Escalate. Speak to your Oracle Sales Rep. Ask to speak to a product manager. These are the ways to get your problems resolved. Believe me, checking email and waiting for news for months on end will not work.
Obviously there are good people and not so good but believe me, I worked with support staff at Oracle who are among the best technical people to be found in the industry. People who really know their stuff. But Oracle has always been a "new features" oriented company, not a legacy support company - a strategy which for better or worse has killed off many competitors in the database market. This was part of the founding ethos of the company and I don't think it will ever change. A decision to move a developer from developing new features to fixing bugs is never made lightly.
So I don't think RedHat needs to worry directly about Oracle providing support for Linux. Oracle will only provide the limited support required to ensure that Oracle products can work properly on that platform. They aren't going to be helping users to set up squid proxy or mysql databases! However, if RedHat's business model is geared around providing premium support for Oracle platforms - then I'd start to worry.
Mostly this concerns me from a personal point of view. I would call myself an "integrator", in so far as what I do is work very closely with HP and HP resellers in pre-sales, delivery, installation and integration of the solution with the customers existing environments. In my experience the customers approach (or are approached by) HP or their Resellers directly, and where Oracle is involved (80% of the time) their sphere of influence extends to the DB implementation and related business, with only some input (for sizing, performance, etc) to the design of the architecture the DB will run on. The rest of the architecture is scoped and delivered through HP working with their Resellers and partners (like me). The choice of hardware and the OS the DB will run on is not decided by Oracle, and they have little or no say in the matter. That's decided between the customer and HP/Reseller in pre-sales technical workshops.
In a "new world" where Oracle can push Son-Of-Raw-Iron, and where the Linux component gives them a huge range of hardware support to choose from and a much bigger mind share grab because of the "Linux" brand image, the whole process of Sales delivery would change. Resellers would be forced to re-evaluate the nature of their relationships with vendors like HP. Perhaps choosing first to take their customers to Oracle before deciding which hardware vendor to get involved. A decision that Oracle would be in a very strong position to influence.
The nature of my relationship with all the parties involved would probably have to change too, as the work I currently do for HP and their Resellers would probably be picked up by Oracle. I'd either have to get an "in" with Oracle, get squeezed out, or be left picking over the scraps.
Well that's one possible scenario anyway. And I might be being overly skeptical, but I have to have a plan ready for that just in case.
Besides the obvious MySQL/commodity low end stuff, you might want to look at specialty high end alternatives too. I have my shoes off as I write this, so counting SAP's BI Accelerator customers is no problem for me; still, that's an interesting product heralding an interesting trend. And I think DATallegro will be on brandname boxes soon too.
My backup for these opinions can be found at various places on http://www.dbms2.com/
To err is human. To forgive is good system design.
Disclaimer: While I am a RHEL customer I have never actually contacted RedHat support so I can't speak directly to the quality level of their support.
I bet this has more to do with adding revenue stream than RedHat support being questionable. Since they plan on supporting RedHat why not take a pot shot (See FUD) at the vendor to help rally customers.
It would seem to me that if Red Hat's support was as bad as oracle is alleging then they wouldn't be kicking ass and taking names in this survey: http://www.bizjournals.com/triangle/stories/2006/0 1/09/daily6.html
How hypocritical, Oracle's standard response to telling them there is a bug or problem with there software is...
"Oh.."
And when asked for a patch...
"Don't know..."
Do oracle actually believe they have good support? Or am I the only one who is getting seriously stressed by there inability to support their products?
Oracle already "supports" any Linux kernel, so long as the kernel complies with Oracle's requirements. But they only go so far as allowing a customer to open a support request (SR) for an obvious kernel issue. If you have issues with anything in userland then you're on your own. Of course if you modify your kernel, Oracle will not spend any time helping you with kernel issues. Will any of this change when they start "supporting" Red Hat?
"Crude and slow, clansman. Your attack was no better than that of a clumsy child."
What a fscking hack!!!!
I am SO going to miss OpenVMS...
"I don't know, therefore Aliens" Wafflebox1
"We can just take Red Hat's intellectual property and make it ours, they just don't have it."
Um.....no you can't Larry. Redhat owns its name. It owns its logos. It uses these to maintain its control over their OS. You can't just take Redhat and stick it in a box and sell it as your own. Redhat will sue you into the ground for using their trademarks without permission. Just like Balmer ("the GPL is viral!") it seems the leaders of the old guard in the software industry don't get how these things work...
With this move Ellison is making Redhat's name (which he does not control) more valuable. That means more money for Redhat in the long run. One again he is trying to rule the world but ends up shooting himself in the foot. That is what he and you are missing. If I was a Redhat today, I would be popping corks...
Open Source Sushi
The Hot-Potato TARs on the Apps side drive me nuts (this isn't inventory, it's OM. Transferring. . . This isn't OM, it's inventory. Transferring . . . ). In fact, I'd say I'm a happy camper as a consultant on every project until I get the first "annoying" TAR, and then I'm trying to figure out what I really want to do with my life.
They have 2 TARs open right now for one client where if Oralce Support submits another response like the last two, my client is going to stop considering dropping Oracle for a custom solution, and instead do it, careers and costs be damned.
So now I get to spend my days escalating TARs up through various duty managers to find someone who can save this client. Wheeee.
You better watch out, there may be dogs about . .
Why miss it? It's been ported to Itanium so its not going away any time soon, and despite the slating Itanium gets from rags like The Register, they're actually very good systems. Very solid and very quick. I've replaced a number of ES45's with rx4640's and the latter whoops the former on performance. It seems to be a much better balanced system for I/O too. Plus Montecito is what, a month away, which will make Alpha OpenVMS upgrades to Itanium even more attractive.