Oops! True IT blooper #97: DOS, we hardly knew ye
by David Pye
The easiest way for seasoned IT vets to date themselves in this world of young programming prodigies is to begin pining for the golden days of DOS. In this blooper, one such codger recounts his frustrating dealings with a pesky .pst file -- and tells how the beloved Disk Operating System of yore would have been able to solve his problem quicker than a Model T Ford rolling downhill in a windstorm.
| S P O N S O R E D B Y : NetIQ |
|
|
Free manageability eBook from NetIQ
Got systems management headaches? We've got your cure -- read our free eBook,
"The Definitive Guide to Enterprise Manageability." You'll learn how to
get enterprise-class event and performance management of your Windows and
mixed IT environments.
Register Now!
|
|
Our lovable 36-year-old coot begins his tale by describing a situation common to many Exchange administrators. As we all know, without frequent clearing out of the 'Deleted Items' folder, the corresponding file can balloon very quickly. "I ended up with a .pst file on our server that had grown to over 120 MB in size," he said. And I'll bet he also had to walk to school through 10 feet of snow, with barbed-wire wrapped around his shoes for traction.
If the sheer size of the .pst file wasn't daunting enough, it also contained more than 1,700 fragments. Our crusty old admin attempted the obvious; he defragged the entire server with a leading utility and sat back in his rocking chair to await the results. They were not good. "Not only did it not defragment the file, but the number of fragments actually grew -- and only in this one .pst!" He was beside himself. Back in his day, this sort of thing just wasn't tolerated.
Methuselah then moved the file off of the server, defragged it again with the same utility and replaced it, thinking he might get a different result. But let's face it, folks, if it had worked, this just wouldn't be a blooper.
Genuinely perplexed by the unexpected increase in fragmentation, he then ran a few tests of his own design. "I copied the file to and from different disks on the same machine, to and from different folders on the same drive and to and from different machines altogether." This exercise, however, wasn't very enlightening. The numbers just didn't make sense. If only he'd had a slate and some chalk handy to help him with the mysterious file size 'rithmetic.
Our wizened friend then decided to compare copy speeds. "The difference between fragmented and unfragmented times was so marginal that it was astounding," he said. After five trials of each, he realized that "the fastest fragmented version was actually moving quicker than the slowest unfragmented one." The average improvement was less than 2 seconds, and Matlock was forced to request a recess to gather his thoughts.
While modern applications do indeed technically defragment your hard drive, do they really increase system performance all that much? Our grizzled friend doesn't think so: "All the 'independent' test results are presented by people trying to sell something, which reduces their trust value to zero." Whether he's looking back through rose-colored glasses or his memory is getting foggy in his old age, Gramps would still replace today's applications quicker than his hip after a bocce accident. "In the old DOS days, defraggers produced a dramatic increase in speed," he said. "Nothing I've tested recently comes close."
So what's the moral to this age-old yarn? There really isn't one. Our hero was shipped off to Shady Pines without ever finding out why his .pst file could not be tamed. But he did not go gently into that good night. You see, he knew a thing or two about a thing or two, and left his whippersnapper IT counterparts with a final message. "I have yet to see a documented defrag test method that led to a reproducible improvement in performance," he said. "I challenge a company with no ax to grind to make the effort to develop a test method and post the results."
There's no certification in "crotchety," but the Rip van Winkle of defrag makes some interesting points that span the generations. Is the practice tangibly useful or simply a marketing concoction? Hopefully, we won't have to fall asleep under a server for a hundred years before we find out. Let us know what you think; e-mail us at [EMAIL PROTECTED].
|