From: Daniel47@teranews.com   
      
   Aragorn wrote:   
   > On Thursday 22 November 2012 15:49, Daniel47@teranews.com conveyed the   
   > following to alt.os.linux.mandriva...   
   >   
   >> Aragorn wrote:   
   >>   
   >>    
   >>> So far, I have found that with Mandrake (prior to the namechange),   
   >>> PCLinuxOS and Mageia, the installer allows you to check for bad   
   >>> blocks but only does this - or so it is my understanding - for the   
   >>> partition designated as swap.  This is not to say that it cannot be   
   >>> done - or perhaps better worded "that one cannot ordain a bad blocks   
   >>> check" - on any partition, but the fact that the installer only   
   >>> offers the option of checking for bad blocks on swap partitions -   
   >>> again, if my understanding is correct - suggests that it's a pretty   
   >>> futile time waster.   
   >>   
   >>    
   >>   
   >> Aragorn, I've got no idea really, but could it be that whilst   
   >> installing a system to a particular partition, that partition is   
   >> subjected to a "bad blocks check" automatically, but, as nothing is   
   >> normally installed to the swap partition, it would not, otherwise,   
   >> have a "bad blocks check" done on it, except at this system   
   >> installation time??   
   >   
   > Short of witchcraft, the only thing that could be checking for bad   
   > blocks more on parts of the hard disk which are being written to is the   
   > hard disk itself. ;-)   
   >   
   > All modern hard disks have a built-in bad block remapping algorithm, and   
   > this is executed on the fly whenever a bad block is encountered.   
   >   
   >> So, if the "bad blocks check" of the swap partition was not done, the   
   >> bad blocks may never by located/marked, and data might be lost??   
   >>   
   >> Just asking/suggesting.   
   >   
   > I rather suspect that, while bad blocks on a data partition may cause   
   > all sorts of mayhem, bad blocks on a swap partition may actually cause   
   > memory corruption and as a result, a crash of a significant userspace   
   > component of the operating system.   
   >   
   > I'm just wagering an educated guess, mind you.  Either way, checking for   
   > bad blocks via software is futile on any first generation IDE hard disk   
   > and later - and this includes SCSI as well, because IDE stands for   
   > "integrated drive electronics", and SCSI disks are thus IDE disks, even   
   > though they are not ATA disks - unless you're only doing it as a way to   
   > see whether your hard drive is on its way south, and even then still,   
   > you wouldn't get this kind of information via a GUI partitioning tool.   
   >   
   > Just skip the bad blocks tests and learn to rely on the hardware's   
   > built-in monitoring.  It's smart enough [*].  That's what it was   
   > designed for. ;-)   
   >   
   >   
   > [*] And it may actually even be S.M.A.R.T. too. :p   
   >   
      
   O.K., was just suggesting..............   
      
   Daniel   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)