Re: CD-ROM hardware failures on 2.4.23-pre5-ben0
On Thu, 2003-12-25 at 18:28, Martin-Éric Racine wrote:
> Basically, the kernel routinely tries to access the CD to check something. This
> is progressively wearing down the drive, as there is often nothing to be mounted
> (drive empty) yet the kernel keeps on trying to spin the drive up.
>
> This problem also existed on previous kernels.
>
> Lately, this drive wear has started affecting normal CD operations, because the
> motor is no longer able to keep a stead pace for long, it starts hickuping. In
> essence, the kernel's needless continuous drive parsing is killing the hardware!
>
> Is there any way to make the kernel stop that routine parsing and instead have
> it try anything ONLY when a media gets mounted or when the CD-Player tool plays
> an audio CD, otherwise leave the drive alone?
>
> Another problem (much likely related to the same bug) is that if I don't leave
> any CD (audio or data - same result) in the drive, the kernel eventually reaches
> the conclusion that the drive cannot be accessed and the only cure is to reboot;
> neither audio or data CDs can be accesed anymore, I instead get IDE errors.
At least 2.6 kernels don't seem to do any of this here (I'd hear if they
did, wouldn't I?), I don't recall 2.4 kernels doing it either. Have you
enabled any automounting related options in the kernel? Or do you have
any automounting related daemons running? (magicdev seems to work
without constant activity here though)
--
Earthling Michel Dänzer | Debian (powerpc), X and DRI developer
Software libre enthusiast | http://svcs.affero.net/rm.php?r=daenzer
Reply to: