Marc G. Fournier wrote:

Morning all ...

  Today, I installed a new Intel server with a 2120s RAID controller in
it, with the 4.7-RELEASE CDs, only to find out afterwards that there is a
bug (known) in the driver causing it to TIMEOUT on reboot ... know, I know
that Scott knows about it, and is looking into it, based on a posting on
Jan 16th to this list, so this question is two part:

1. has it been fix in recent code?  Will an upgrade to -STABLE (which is
   planned anyway) fix it?  I found a commit since the bug was reported,
on the 23rd ... is this meant to address that problem?

"Work around a bug in the 2200 and 2120 cards that cannot DMA commands
  from the first 8k og physical memory."

Yes, this checkin addresses the last known problems in the driver.  If you
can update to 4-STABLE, the driver will work.


2. I doon't know how the Hardware RAID controllers work, but is there an
alternate card that I can put into the machine (ie. the 2110S (U160 vs
U320) that will work, or is this an issue that affects anything using the
aac driver altogether?

To my knowledge, it only affects the 2120 and 2200 cards.  There really
isn't a good reason to switch to the 2110S now that the bug has been fixed.

Also, how does hardware RAID work?  With vinum, I know that the 'RAID
configuration' is saved on the drive itself ... where is it saved on a
RAID controller?  On the card itself?  If I replace the controller, do I
end up losing all of my configuration?  Or is it simple 'plug in new card
and reboot'?

The 'metadata' is saved on each drive that composes the array. It is also
saved in nvram on the card. When you turn on the machine, the card scans
its scsi busses and compares what it finds to what it has stored, and
configures itself accordingly. If you replace the controller, the new
controller will scan the buses, see the drives and arrays, and configure itself
to use them. It's designed to be pretty seamless and logical in this respect.

Scott



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-stable" in the body of the message

Reply via email to