Volker Kuhlmann wrote:
I did a blockdev --setra 8 /dev/hdd as root ( 8*512 byte = 4k ).
Unfortunately the problem remains the same.
It is necesssary to first test whether you're running into said kernel
bug, before considering other possibilities. For this, you will need to
set read-ahead to zero, not something small.
hdparm -a0 /dev/hdd
will do it. So might blockdev. Check with hdparm /dev/hdd
AFAIK different layers. hdparm works on the drive, blockdev may work in
the o/s, since I can ret readahead far larger than the drive claims to
support. And I believe on CD/DVD blockdev is sufficient, it works for me
when verifying copies.
However, by all means set or check it both ways, I haven't waded through
the code.
Note this will make your DVD drive very slow. It is sufficient to
disable read-ahead from another shell while your reading is going on, as
long as you make sure that you do it well before the kernel reads near
the end of the recording.
What next ?
Provide missing info ;)
Volker
--
E. Robert Bogusta
It seemed like a good idea at the time
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]