Approach.Botonomy.com has an interesting article on Why the Puritan Work Ethic has No Place in IT.

The article echoes many of my personal sentiments that your administrators and developers need to be encouraged to work smarter, not harder, even to the point of having “free time.”

… You want an environment where sysadmins kick back and read IT magazines occasionally, because their run-of-the-mill administrative tasks (adding users, managing disk space, etc.) are all scripted and/or automated. They can then focus their energies on the unexpected and unavoidable issues that crop up from time-to-time.

Beyond handling the unexpected, through having this “free time” administrators will have the ability to identify areas in need of improvement. If your administrators are running around fixing stuff all the time your team has a problem! Not only will morale, and therefore retention suffer but your administrators will have no time for evaluating new opportunities.

The best teams celebrate those who sit back and let their computers do their work for them. You want to have a project team that considers repetitive development activities to be tasteless. Sometimes necessary, but generally frowned upon.

Check out the full article and think a bit about what your team could be doing if they weren’t fighting fires all the time.

Of course this is exactly why I am writing a book on Oracle Shell Scripting due out next year.

project management, it management, information technology, database administration, system administration

Easy Linux Commands: Working Examples of Linux Command SyntaxI’m happy to announce that my book Easy Linux Commands: Working Examples of Linux Command Syntax can now be preordered from Rampant TechPress!

This book by Terry Clark and I is designed to serve as both a progressive how-to for getting comfortable with the Linux environment and as a reference for the most commonly used Linux commands and options. After presenting a little background information the books jumps right in to how to navigate the Linux file structure, manipulate files and directories, search for specific objects and get around in the vi text editor.

More advanced topics like shell scripting, using cron to schedule tasks and monitoring and administration tasks are also covered all with the focus on giving working examples of the Linux commands you will find useful. The command examples in this book can often be used with little or no modification saving considerable time and experimentation.

This book organized into logical task-based chapters making it easy to find the commands you need when you need them.

For a full table of contents and index check out the book’s page on Rampant TechPress. It’s available for preorder through Rampant and should be in book stores in just a couple months!

linux, linux commands

Apple's Xserve RAIDIt looks like Oracle has decideded to adopt Apple’s XServe RAID as a low-cost storage solution

Based on our own experience with Apple technology, Xserve RAID is a great match for applications running Oracle.

With the appetite of one to two petabytes annually Oracle is of course looking to control costs and the SATA-based Xserve RAID combines value, capacity, performance and availability. The current top configuration offers 7000GB on dual RAID controllers with redundant power and cooling for a mere $12999. Sure that’s the price of a small car, but it works out to only $1.86/GB for some serous industry-level storage.

Thanks to Zach for sending this on to me.

apple, oracle, xserve, raid, xserve raid, storage, disk, hard drive

While researching the effect of RAID 5 disk configuration on Oracle databases I came accrost more than I thought I expected on the topic.

With disk as cheap as it is today there is no reason to ever use RAID 5 on an Oracle database. Even for a database which sees only a moderate amount of updates the performance loss incurred by using RAID 5 is too high. RAID 1 or some combination of RAID 1 and RAID 0 should be used when high availability is required. Database 10g users should also consider allowing Oracle’s new Automatic Storage Management to handle redundancy on un-mirrored disks (or even raw partitions.)

But don’t take my word for it. Here are some opinions from some notable Oracle administrators:

Oracle Database RAID 5
Mark Rittman

No RAID5. Use RAW whenever possible and consider RAID10 or 0+1. RAID 5 can severely affect performance on highly updated databases.

Oracle Database Administration: The Essential Reference
David Kreines & Brian Laskey

RAID-5 is, in fact, very powerful and inexpensive. It is also a technology to be avoided in most cases when configuring your Oracle database! This may seem a harsh statement, but the reality is that although RAID-5 provides good levels of data protection at a low monetary cost, this comes at a very high cost for disk I/O. In particular, write operations on RAID-5 arrays can be orders of magnitude slower than the same operations on a single disk.

Oracle and RAID usage
Mike Ault

Use RAID10 when possible, RAID5 if it is not. Size the array based on IO needs first, then storage capacity and you can’t go wrong.

Disk Management for Oracle
Donald Burleson

Oracle recommends using (SAME) Stripe And Mirror Everywhere (a.k.a., RAID 1+0) for all systems that experience significant updates. This is because of the update penalty with RAID 5 architectures.

Using RAID 5 for a high-update Oracle system can be disastrous to performance, yet many disk vendors continue to push RAID 5 as a viable solution for highly updated systems.

RAID 5 is not recommended for high-update Oracle systems. The performance penalty from the parity checking will clobber Oracle performance.

If you’re still not convinced check out Baarf.com. This site is committed to the Battle Against Any Raid Five (or four or free, uh, three.) The site has links to several more articles exposing the perils of RAID 5.

oracle, rdbms, dba, systems administration, sysadmin, raid, raid 5

Sun T2000Well things weren’t quite as easy as Johnathan Schwarts, now CEO of Sun implied when he wrote “if you write a blog that fairly assesses the machine’s performance (positively or negatively), send us a pointer, we’re likely to let you keep the machine.” but after a little extra paperwork and a more formal contest we have won a Sun Fire T2000 server!

That’s right! Thanks to considerable help from Scott Maziarz, an outstanding Plymouth State University IT major who has been working with me on all things Oracle, we were able to evaluate the performance of the Sun Fire T2000 against our existing Sun Fire V440. Scott ran a series of tests based on some jobs we run daily to recreate data marts in our production system.

Scott ran the tests and compiled the results and I wrote up our findings here on Life After Coffee for others to see (and so we could enter the contest.) We ended up shipping our T2000 back as our 60 day trial was over before we were chosen for the contest, but now Sun says they’ll be sending it back to us.

So a big thanks to Scott for his help and to Sun for running this program and choosing us as a winner!

« Previous PageNext Page »