Re: [cobalt-users] df and updatedb result in proc state of "D"

[flash22@xxxxxxx]

On Sun, 31 Mar 2002, Mike Jeffers wrote:

> I'm finding that if I manually try to run "df" or "updatedb" the processes
> will INSTANTLY go into that state and freeze.. Manual pages define the D
> state as:
> 
> D   uninterruptible sleep (usually IO)
> 
> 
> This leads me to believe that I *may* have some file system corruption or
> problems, but I don't have any hard evidence to back this theory up with.
> I'm very concerned about the stability of my machine right now.

Yes, run badblocks , see if you have a media flaw on the disk, dmesg
should also log an system level errors (i/o errors, VFS corruption) 
[dmesg is a program that prints out the last few kernel messages]

since you seem to be able to log in at all, it's probably not running out
of swap, but you could have a bad block there as well..

D = Demand paging (obsolete term basicly means it's waiting for a page to
be swapped into memory somewhere, why it's waiting it doesn't tell you
tho, 'Z' is far far worse ;P]

Rebooting will make the machine fsck the filesystem, if it's a minor
corruption instead of a media error it should fix it, but  it also might
not come back up if it's really screwed up...

[hard drive reliability for high capacity drives seems to be getting worse
these days, sigh]

gsh