Hello, friends,

I work with audio and video recording, which demands high sustained transfer rate from Hard Drives. I´ve noticed that some drives, although without bad blocks, have some slow clusters, which cause trouble in multitrack recording.

I wonder if there were a method to scan all HD and isolate all slow sectors to a place in which I could confine them to a "dead partition".

MHDD software has a function to erase brown and red blocks by writing to them. But does it really isolate those blocks ? or simply try to write on them ?
I wouldn´t mind if I could take my 200Gb Seagate drive, which have some troubles, and turn it on a 100Gb REALLY FAST drive, less than 50ms, if MHDD or another software could automatically scan al disk and move slow sectors to another place.

Just an Opinion: Software Industry is too much worried about recovering Bad Blocks, even when they will be very slow. I don´t need slow hard drive!
Software SPINRITE for instance, try and try to read sectors and finaly declares as GOOD !!!

No way! I think that, in hard drives, if the first time you can´t read a sector, software should mark as bad block, so we wouldn´t ever have a slow drive in hands.

Any hints are welcomed!

Thanks,

Dihelson Mendonça

i would sell old drive on eBay saying exactly what is wrong with it, or just chuck it out, and buy a new one, they are cheap enogh these days, not worth stressing over this, once it starts waring out, just gets worse from now on, not better, just my opinion

Mhdd and spinrite doesn't work for you. Defects should be isolated with selfscan or working with err lists, but need expensive gear to do it. Buy a new drive, and when you have some drives, give it to a pro for repair - won't cost that much.

Buy a new drive.

_________________
www.datarecoveryne.com

OK, let's see what you really want.

May I suggest SCSI and SAS for ultimate performance if you want a drive system capable of sustained heavy I/O at very high rates of acquisition?

While 10K RPM is cool, 15K RPM is fun stuff! You can get this level of technology in used Dell servers for a song along with a pro-level RAID controller and the SCA backplane as part of the package.

Failing that, consider single-platter current generation drives. For example, for Seagate 7200.11 series, that would be a ST3160813AS 160GB drive. According to specs, the 640GB drive is the fastest, however, as the write time is improved from 10msec to 8msec.

Those numbers become laughable when you look at true performance: 4.3msec on WD VelociRaptor drives, less than 4msec on Seagate Savvio and Cheetah drives (priced extremely high).

One of the fastest drives I'd ever scanned in MHDD in terms of slow sectors was a Seagate 7200.8 series drive. I couldn't believe my eyes at how clean the scan was. I think it was a 200GB unit.

For your application, a couple of WD VelociRaptors in RAID0 would probably yield the best price/performance ratio. If you wish to get that level of performance but without the risks of RAID0, get four of them in a RAID 10 array on a battery backed cache hardware SATA RAID controller. Such an array would sustain the failure of one or two drives, as long as they are not part of the same mirror and let you replace the failed drive as a matter of routine without degrading performance. Avoid RAID5 for data acquisition systems. RAID5 is fast for reading, but not ideal for heavy write applications. When it enters degraded state, the performance tanks.