[sclug] Cheap low-end servers

[lug at assursys.co.uk]

On Sat, 4 Oct 2003, Simon Heywood wrote:

> The Red Hat installation program reckons there are bad blocks on the
> drive (but won't tell me where they are) and advises me not to use the
> disk. However, an e2fsck bad block (write) test didn't pick up anything
> on the first run (I had to power the machine down so didn't see the
> whole thing finish). Any suggestions...?

badblocks normally only does a read-only, non-destructive test, so I'll
assume from your '(write)' annotation you're using the -w switch.

IME, some (all?) discs only remap failed blocks on write, so writing all
over the disc (i.e. /dev/hdx or /dev/sdx) should make it all OK again (until
the next block fails, or the disc runs out of spare reserved blocks). I've
"fixed" partitions containing bad blocks by deleting the file that occupied
the bad blocks in question, and dd'ing a file to fill up the partition (e.g.
dd if=/dev/zero of=/var/foo bs=512).

If a write test over the entire drive has completed successfully, try the
install again. If it still doesn't work, chances are the disc has run out of
reserved blocks and should be binned or returned. If it does work, continue
to monitor the health of the disc using
<http://smartmontools.sourceforge.net/> or similar.

> S.

Best Regards,
Alex.
-- 
Alex Butcher      Brainbench MVP for Internet Security: www.brainbench.com
Bristol, UK                      Need reliable and secure network systems?
PGP/GnuPG ID:0x271fd950                         <http://www.assursys.com/>