Pythian Group

This is the 137th edition of Log Buffer, the weekly review of database blogs. Dave Edwards is enjoying a week off, and so as part of my plot to take over the world, I am writing this week’s Log Buffer.

First, the fun stuff: Josh Berkus tells us that the American English Translation of the Manga Guide to Databases is available in Japanese Fairies and Third Normal Form.

Then, the basics:
Giri Mandalika points to an article on Using MySQL with Java Technology. This is a basic article on how to connect, and does not go into all the wonders that Connector/J can really do. Speaking of basics, Decipher Information Systems has an article on Back to the Basics: Refreshing a Regular View Definition in SQL Server and Oracle. Chet Justice, the Oracle Nerd, writes about transactions in ORA-8177 Can’t Serialize Access For this Transaction. There is theoretical information about isolation levels that applies to all database systems. Mats Kindahl explains some transactional theory when talking about why Mixing Engines in Transactions in MySQL can get you into trouble.

It’s hard to get more basic than what data type to use for a single character Read the rest of this entry . . .

A turning point in a DBA’s career is when he/she is recognized by Oracle as an elite Oracle professional. I was happily surprised to find out I became an Oracle Certified Master (OCM). I got another surprise right away, “You are the 2nd OCM at the Pythian Group”. Well, I am very happy that I have been working among brilliant people here.

Believe it or not, after I became an Oracle 10g OCM, everybody else wanted to know how to become one. Let me show you the way to become an Oracle Certified Master (OCM).

The practical way to become OCM

1) Be Oracle OCP first. If you don’t know what that is, stop here.

2) Take 2 advance trainings from Oracle University. I highly recommend RAC and SQL tuning. Taking the training, I can play with RAC at my will without any concerns. RAC in the Oracle lab doesn’t belong to any production instance, so no worries at all. Furthermore, refreshing your knowledge of tuning is always good. Your experience from work only can’t substitute as the full toolkit of Oracle — which is what you are tested on.

3) Register for the OCM test (in many cases, the test will be in another city)
It is true for all Canadians. I had to go to New York to take the test. I was told by Oracle that there is no plan to provide an OCM test site in Canada. And I met another Canadian DBA in the test, from Toronto.

4) Book plane and hotel then take 2 days to become OCM
Get a hotel close to Oracle building though it is probably expensive, let’s say CAD $200 a day without tips. Don’t get lost in train/substation during rush hour — get a hotel close enough to walk to the exam.

5) Get results in 1 month.
I got my ‘congratulations letter’ from Oracle in a month, then a T-shirt some weeks later with the OCM logo. In this step, just wait in your warm home and enjoy it.

Drop me email if you still have questions. Welcome, my DBA friend.

I’m not sure how long this has been out there, but there is a new (to me) headline on Oracle’s support website, announcing that next month, they will be phasing out “manual configuration” information for service requests.

Customers are now required to download and install something called Oracle Configuration Manager (OCM), which will gather their system/database configuration information automatically, and forward it to Oracle Support on their behalf.

I don’t know a lot about this tool. Yet. The OCM page on on Metalink offers the following description:

Configuration Support Manager allows you to define computer configurations that describe your Oracle environment, and milestones for projects involving Oracle products. Providing this information will allow you to log SRs with less data entry, track issues more effectively, and will allow Oracle to proactively suggest solutions and resolve issues faster

OCM has been around for a while now, and doubtlessly it does have its uses. For example, I was recently looking for an easy, platform-independent, way of determining the latest CPU update applied to an Oracle home, and OCM popped up right at the top of the list. But, like most DBAs, I’ve been far to busy actually managing databases to really give this tool much thought. Until now, anyway, since it seems I (and you) will have to start using it really soon.

Based on a cursory look and on a few conversations around the office, I gained the impression the OCM is little more than a sophisticated piece of spyware. “Services” like Windows Genuine Advantage spring to mind. You know, “services” that compel you to submit to a scan of your system to establish the validity of your licenses before you can obtain updates or service, and perhaps even threaten to do “bad things“^TM if the scan happens to fail.

I find, though, that upon starting to actually read the OCM documentation (imagine the desperation that motivated that!), OCM does not look quite as scary as it did at first. It still creeps me out somewhat — my vivid imagination has no trouble conjuring up unpleasant images of what Oracle might do with this data once they have it — but the documentation I have read thus far has at least tempered my concerns. Here is what I have found so far:

• Contrary to suggestions I have found elsewhere on Metalink, OCM does not have to report data directly to Oracle support. It does do this when operating in connected mode (the default), but it can also operate in disconnected mode, where collected data is deposited in a .jar file that you can (I hope!) inspect yourself before electing to forward it to Oracle. The availability of disconnected mode makes OCM at least feel less like spyware.

Okay. That’s a short list. But I’ve only just started. I’m sure I’ll find more, right?

Anyway, the option to run in “disconnected mode” makes a fundamental difference here, as it returns control of the conversation from the software vendor back to us, the customers. Providing the data collected is not encrypted (or so horribly obfuscated as to be completely useless), this means that I have the ability to review what is being collected, and decide for myself whether or not I choose to disclose that information and how much I will disclose, or even reuse the collected data for my own purposes.

So, why the big deal? Why am I concerned about letting Oracle support see the “unvarnished” truth about my system configurations? Well, in all honesty, I do find myself telling Oracle support occasional “little white lie”. By nature, I am a very truthful person, but I can imagine legitimate (or at least justifiable) reasons to withhold certain details from Oracle Support. For example, when opening simple support requests, I can understand how a person could comfortably conceal the fact that their test environment runs a binary compatible clone of a supported operating system, rather than the (much more expensive) commercial release. In situations like that, Oracle Support could (okay, maybe legitimately) deny that person service, even though though the problem has no conceivable connection to the fact that their database runs on CentOS rather than Redhat Enterprise Linux.

On the other hand, OCM is a tool that shows considerable potential, much like RDA. I have already imagined a few internal uses for data collected by OCM, but I won’t be writing about those until I take the time to verify that OCM actually collects the data I am after and lets me see it. At the very least, OCM does promise to make it easier for me to open an SR with Oracle support, by removing, I hope, a lot of that routine dialogue I spend 15 minutes completing every time I open a support request.

So, now I find myself completely ambivalent. Or perhaps conflicted. Or maybe simply confused.

Does OCM represent some sort of Orwellian conspiracy, or is it an incredible blessing, unsought and unexpected? I just can’t tell right now. Could it be both? Ooh! My head is starting to hurt just thinking about that one…