On Wed, 19 Jul 2000, Henry Nelson wrote:

> I'm trying to install apache from the  package, but I got an error code 1:
>    mkdir /usr/local/lib/httpd
>    sd0(esp0:3:0): medium error, info = 1955972 (decimal), data = 00 00 00
>    00 11 01 00 00 00 00
>    mkdir: /usr/local/lib/httpd: Input/output error
>    *** Error code 1
> 
>    Stop.
> 
> This is on a Sun 4/40 (sun4c), NetBSD 1.4.2/sparc.  What's different from
> what I've done before is that LOCALBASE is set to /usr/local, which is
> mounted on a different partition (sd0f) from what /usr, including pkgsrc,
> is mounted on (sd0d).
> 
> Should /usr/pkgsrc and LOCALBASE be mounted on the same partition, or is
> there something wrong with the disk?  The disk is a rather lightly used

	Its almost certain a disk hardware issue, or maybe a cabling or
	controller issue (less likely).

> Quantum Fireball TM1280; there's never been an indication that it was
> bad before.  Does NetBSD have standard tools for scanning a disk to look
> for errors, and then mark any bad sectors?
> 
> What's the recommended way to proceed in such cases?  TIA
> 
	I've tended to backup, run 'scsictl sdX format', and reinstall,
	but you should be able to locate the bad sectory by just dd'ing
	from the raw partition, then use 'scsictl reassign NNN'


                David/absolute
			       -- www.netbsd.org: A pmap for every occasion --