Joel Fowler wrote:
> Brian Haymore wrote:
> >I have used standard and the UDMA/66 cables and they both get the same
> >performance. The drives are not able to sustain more then about
> >12-18MB/sec on block IO.... They can burst to bus speed(33 or 66) but
> >not sustain it. No vendor even claims this...
>
> My information (I assume) reflects a sustainable transfer rate
> between disk buffers and central memory. We both know there's
> a lot more to disk I/O than buffer transfers.
True that max disk transfer rates such as 33MB/s for UDMA only happen out of
the drive buffer. But I think what Brian was trying to say is that there is
no speed difference between udma-66 cables and normal eide cables.
I use a mix of udma-66 cables and eide cables (udma66 for longer runs) and
can verify that there is no speed difference. The identical drives on the
identical controllers run no slower or faster (burst or sustained) when
attached with udma-66 cables. But they let me go long, woo-hoo ;)
> There seem to be so many variants of software raid configurations that I
> find it difficult to determine what documentation/advise applies to me. I
> wish there was a "Migration to Full Software Raid-1 with Recovery
> Scenarios HOWTO".
This is a pretty helpful document that was posted to the list:
http://kernelnotes.org/lnxlists/linux-raid/lr_9904_04/msg00032.html
Be sure to read the follow-ups.
I don't think there is a document that describes failure types and recovery
scenarios, however much of it has been discussed on the list in the past
months. Try:
http://kernelnotes.org/lnxlists/linux-raid/
To peruse what people have said.
Good luck,
Tom
