Hi,If you are like me you have been living a life of permanent anxiety as you wait for your HD/s to fail :(Now I know most are pretty reliable these days, but that is small comfort when you hear tales of unexpected failures with loss of hours/days/months/years of work.I imagine the majority of us have mainboards that are S.M.A.R.T capable and enabled, but can we really rely on this monitor to warn us in plenty time before disaster strikes?This may well be the answer to our prayers:http://www.passmark.com/products/diskcheckup.htmI, for one, am feeling much better now! Both my HDs appear to be in tip top shape and regular checks using this utility should give me all the warning I need to act before it is too late. I still have one little niggle, however. Can this utility predict an impending mechanical failure?Mike :)

Thanks.And for Linux people, I use Webmin, which has SMART info as well.

Mike,>>Both my HDs appear to be in tip top shape and regular checks using this utility should give me all the warning I need to act before it is too late.<

Hi jbNo offense intended, but that's livin' in a fools paradise. Your little niggle sez it all.....none taken ;) You're probably right, which is why I have been taking regular complete system backups for a couple of years now. You really can't be too careful if you value both your data and complex installations such as FS9 which have often been carefully built up over an extended period of time.Still, mechanical problems aside, this useful little utility does have some value and should serve to warn of a deteriorating trend in the performance of your drive/s. A slide towards the threshhold values could well be indicative of an impending mechanical disaster and will signal the need to transfer everything to another drive before it either becomes corrupted or it's too late. In the case of the latter scenario all that is left is to restore your latest full backup which, in my case, could be anything up to a month old.Mike

Hi Mike,It amazes me how many people don't backup. Jeesh, even if you do it manually, a days work doesn't take *that* much time. Agree on the trend idea, useful to have a heads-up in advance.Another nice util is HD Tune 2.51 (freeware):http://www.hdtune.com/HD Tune has a good handful of error reporting parameters and indicators if things are goin' south, as well as basic benching. I also use PerformanceTest from Passmark (payware version), its got some nice features, but HD Tune and some other free utils often provide more info . One you *don't* want to pay for is Futuremark's PCMark 05. I'd call it a pos, but that'd be understating the case ;-)jb

Another utility you guys may have heard of is Spinrite.Sounds like it pretty much does the same thing as what you all are talking about here.

Mace,That's from GRC? It just came up on another forum discussing HD durability. I've used Gibson's IDserve and the port testing, hadn't tried Spinrite.jb

Hi.There is only ONE answer to HD problems / failure, BACKUP. You can buy a USB Enclosure for about $20.00 and an 80-100GB HD for about the same amount. Keep your OS on a separate partition and BACKUP your data (use Copy / Paste)/ OS (use Retrospect or Ghost for OS) after every major install or file created. It can take as little as 2 min. a day. Turn off your USB drive when not used for backup and this will also protect you / it, from viruses, assuming you get them before you backup. TV

I don't know if Spinrite is made by GRC. I have not done much research into it, really--but I thought it was similar in concept to what was being discussed. I did read through the general concept behind Spinrite, and it seems like a valid one, although I always question "overexercising" the sectors, which I would think in and of itself may cause excessive wear.

>>...I always question "overexercising" the sectors, which I would think in and of itself may cause excessive wear.<

>His analysis, *as I read it* is that the>extra wear and tear may reduce the physical life of the drive>by some minute percentage, the constant data moving keeps it>'fresh' and therefore perhaps less likely to get lost. Heh that makes it clear as mud. On the one hand he says it does reduce the phys life of the drive but on the other hand the data movement makes the drive less likely to lose data. Kind of a catch-22.So in your opinion, it is a better trade-off to defrag pretty regularly (say, once or twice a week on a light-to-moderately used system) than it is to just defrag once a month or so?You use Diskeeper 10...I was thinking about getting it or PerfectDisk, but I could not decide. Another (myth??) I have read about is that if you use the built-in Win Defrag, you get more complete degragmenting by using it three times in succession. This SEEMS to be borne out by the visual display of fragmented files, but that's hardly a scientific way to analyze it. Have you heard of this?Rhett

>>Heh that makes it clear as mud.<>Balancing that, of course, is the increased disk activity that the defragging causes. Because the system is busy, the drives spin down less often; the bearings are under additional load, and the seeking during defragging is relentless, which causes wear on the disk head swivel bearings and increased heat dissipation because of the voice coil current. It likely decreases the total life expectancy of the drives by some imponderable percentage.<http://www.pcmag.com/article2/0,1759,1811993,00.aspI've always run the built-in 3 times, and it does look better by the third pass, but I've heard it referred to as 'rudimentary defragging'. I think I read that DK technology may be used in the MS defragger, but not implemented as it is in the stand alone. I know that DK makes my machine visibly faster, and that was after an XP defrag. I keep meaning to run BootVis (cool, discontinued MS util, available on the web though). It tells you exactly what's loading, and how long it takes, and can also optimize boot order. Use with caution :-) XP defrags on boot every three days I believe, but you can have DK do that whenever you reboot. I believe it's the MFT, paging file and other sys files that are done at that time. Now I've got a question that's bothering me. For some reason, my WD800JB drive is writing in the high 30's low 40's MB/s range, but it's only reading ~25MB/s. I've tested it with a bunch of utils and the range stays the same. I can mess with the test settings in PerformanceTest 6 and get better results, but that's not the point. It still works great to render mpg's for DVD creation, and actually, plenty fast to capture too, just puzzling. The Dell Seagate OEM drive reads and writes ~ the same speeds (low 40's), the new ST SATA drive is nicely in the 50 range for both. Like I said, more of a puzzle than a problem. It's the Primary slave, UDMA-5, no problems there, SMART doesn't indicate anything bad, wondering whether it's worth making it Secondary Master and putting the CD-RW as slave on that channel. I've also considered making the cluster size bigger, that probably makes sense anyway given the drives usage. Dunno, suggestions welcome.jb

>XP defrags on boot every three days I believe, but you can>have DK do that whenever you reboot. I believe it's the MFT,>paging file and other sys files that are done at that time. >That's right, WinXP's defragger uses technology under licence from Executive Software (Diskeeper). But I hear Defrag is dumbed down like you said. Well I will have to do some more research into aftermarket defraggers, I also hear of one called "O & O" but I don't know what that is yet.>Now I've got a question that's bothering me. For some reason,>my WD800JB drive is writing in the high 30's low 40's MB/s>range, but it's only reading ~25MB/s. I have an old IBM 36 gig 10000 rpm SCSI drive that does that. Write spd was pretty fast but read was not up there. I did not try changing its cluster size. You might try that, depending on what type of filez you push through it. You would think that read and write spd would be relatively close on a hd, assuming the drive isn't badly fragmented, and the data is being written to/read from roughly the same part of the platter. But then I am no hd expert, I have just been around the scene long enough to be dangerous. :) I never could figure out why the IBM drive did that. I still use that drive on my net browsing rig (former FS system).Rhett

Rhett,>> I also hear of one called "O & O" but I don't know what that is yet.<>I did not try changing its cluster size. You might try that, depending on what type of filez you push through it. <. Since the drive is for rendering fairly large .mpg's from TMPGEnc 3.0 Express or its older brother, TMPGEnc 2.524, I'm reformatting again (full format, not the quick deal), w/64k clusters. I fully expect the same results ;-) Glad you mentioned you've seen the same on another drive. I tested an identical WD drive (well, bought the same day, and the S #'s are close) on one of the other boxes, same results.For a SCSI 10k drive to exhibit the same behavior, makes ya wonder.>> But then I am no hd expert, I have just been around the scene long enough to be dangerous.<. Also slung *a lot* of Lingo in Director, through v8.5.1, got paid for it, that was fun. D***, I think I'm gettin' old ;-) Enough nostalgia, the drive is still formatting w/64k clusters, I'll post results. Btw, I'm using PerformanceTest to bench because I still have a *little* bit of a problem setting IOmeter up correctly. Hmmm,, it's the weekend, wonder what I'm gonna do (note to jb: get a life).For lurkers who may still be interested in this thread, a few HD tips....if using an NTFS formatted drive for captures, or some other dedicated purpose, you *may* want to disable System Restore, Indexing Service, Compression, and set the Recycle Bin to 0%, for *that* drive only. Gains some disk space and 3 or 4 CPU cycles. There are plenty of other 'tips' out there, some work, some don't, caveat emptor...well, Google is free, but ya get the idea ;-) Backup your Registry before trying most of em.Edit: With 64k clusters, the same results for the WD drive...~45MB/s write, ~24MB/s read with most utils. Now, if I set the test in PT6 to be:File Size: 1022MB Block Size: 16384 bytesStandard Win32 API cachedSynchronous mode100% Read/100% SequentialWell, I get ~42MB/s reads. Lol, ain't benchmarks cool ;-) Cut the File Size in half and you've got Super Drive (no, not the Mac thing)...jb

Hi all,Well, I liked Diskeeper 10 Home version, then I dnld the trial of PerfectDisk 7.0 (URL below), as my trial of DK was about up. Very nice, and seems to defrag the system files in a much more thorough manner. As I never intended to use the 'Set it and forget it' feature in Diskeeper (constant defragging), PerfectDisk seemed, well, perfect ;-) It also includes network ability, though apparently not for XP Home, though I think that's maybe doable. The other XPPro box apparently needs to be reconfigured permission-wise, I'll deal with that later, (the network deal was never an issue anyway). And, last but certainly not least, Amazon had a deal I couldn't refuse: $31.99US, free shipping and a $20 mail-in rebate. Sold.The salient results from a Diskeeper analysis after boot-time defragging, and *two* passes with PerfectDisk, all non-essential svcs disabled w/msconfig.exe:'Congratulations! There are no excess file or directory fragments on this volume. The files on this volume are as defragmented as possible.'As to the WD read speed issue, I replaced the cable for the Primary IDE channel with a brand new one, hey, no problems, read speed back to ~42MB/s. Shut the box down overnight, checked speed again, back to ~24/25 MB/s, write speeds still in the mid 40's. Rebooted, back to the 40's, I think it's a thermal thing, either on the mobo (most likely) or the drive. At this point, I don't care, it works .jbPerfectDisk site:http://www.raxco.com/