> You reckon that I'm wasting my 
> time with repeated fsck -fc /dev/hda5 then?

Not totally, but it's inefficient on various counts, and ineffective as
far as testing the whole disk is concerned. You can however check
whether all disk blocks are readable with dd if=/dev/hda of=/dev/null.
This also gives you the LBA of the bad blocks(s). Alternatively run
badblocks.

smartmontools is easier and gives you more info. For example if the
number of reallocated blocks increases by one a month it's probably ok.
If it suddenly increases by 10 it's time for a backup. If it tells you
"SMART status FAILED" the disk expects to exit stage left within 24
hours (needless to say you do a backup before going to lunch). The
latter also means nobody will argue about warranty.

I got a total of 3 failed selftests for one new disk weeks after
purchase. Each time I ran the same test again, and it passed. There's
no way normal operation would ever tell you this. That disk was dead 2
months after I bought it. Current versions of smartmontools would have
even sent me 3 emails.

Volker

-- 
Volker Kuhlmann                 is possibly list0570 with the domain in header
http://volker.dnsalias.net/             Please do not CC list postings to me.

Reply via email to