Anders Johansson wrote:
On Tuesday 30 October 2007 15:12:01 Aaron Kulkis wrote:
A filesystem can't get corrupted if there's no write
activity on it.
Of course it can. Underneath it's hardware, and hardware can go bad.
Which usually effects the entire drive, not just one
filesystem on it.
Effects
such as bit flips can cause file system corruption at any time
on a modern disk drive, base level encoding system is
so redundant (error detecting/correcting codes) that the
likelihood of a bit error actually making it up to the
level that an fsck can detect it is, over the course
of a year, is less than 1 in a billion.
Why?
Precisely because a single bit-flip could be so
destructive otherwise.
That's one reason for regularly checking your file systems
The far greater corruption risk is an intermittent
bug in the kernal or device driver creating a
data or meta-data corruption error, or power
failure during a write sequence, or a kernal crash
during the same.
This is why I use journalled filesystems for everything
except for /tmp.
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
