Anthony's drive issues.Re: ssh password delay

Anthony Atkielski atkielski.anthony at wanadoo.fr
Fri Mar 25 13:08:33 PST 2005


Ted Mittelstaedt writes:

> That is right around the time that brand new drives fail, if they are
> going to, that is.

Well, I got a replacement drive today, so if this one fails, I have
another one standing by.  I'll need to see more clear indications that
the drive is actually in trouble before swapping them, however (I have
backups and only /var and /tmp are on the drive, so I can afford to wait
and see).

The self-tests I run with smartctl show no errors, but the UDMA CRC
error count for /dev/ad10 is non-zero, as is the soft error count.  I
don't know how much I can trust these numbers.

> Modern drives with the exception of high end SCSI ones, are as a
> friend of mine put it once: "slapped together a million miles a second
> on the assembly line"

I could buy a dozen of these drives for the cost of one equivalent SCSI
drive.  SCSI is nice, but it's awfully pricey, for no good reason that I
can see, and unless one is running a very heavily loaded server, I'm not
sure that I see the advantage to it.

I was thinking a few days ago that extremely fast static RAM might be
the single best way to boost system performance, but that was just
daydreaming.  (From what I understand, most modern processors spend most
of their time in wait states waiting for memory to reply.)

-- 
Anthony




More information about the freebsd-questions mailing list