From owner-freebsd-current  Thu Sep 23  2:26: 4 1999
Delivered-To: freebsd-current@freebsd.org
Received: from apollo.backplane.com (apollo.backplane.com [216.240.41.2])
	by hub.freebsd.org (Postfix) with ESMTP id 409B915A19
	for <current@freebsd.org>; Thu, 23 Sep 1999 02:25:26 -0700 (PDT)
	(envelope-from dillon@apollo.backplane.com)
Received: (from dillon@localhost)
	by apollo.backplane.com (8.9.3/8.9.1) id CAA23816;
	Thu, 23 Sep 1999 02:25:26 -0700 (PDT)
	(envelope-from dillon)
Date: Thu, 23 Sep 1999 02:25:26 -0700 (PDT)
From: Matthew Dillon <dillon@apollo.backplane.com>
Message-Id: <199909230925.CAA23816@apollo.backplane.com>
To: current@freebsd.org
Subject: CCD update committed
Sender: owner-freebsd-current@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

    I've made a bunch of fixes and enhancements to CCD.  I do not believe
    I've messed anything up but CCD is rather messy so I'm doing a little
    head's up anyway.

    * CCD panic when underlying storage has a sector size larger then 1K
      fixed.

    * CCD lockup when dd'ing from the raw ccd device using a block size
      that causes the CCD partition's EOF to occur in the middle of an
      I/O fixed.

    * CCD complains to dmesg (but does not yet cause a configuration failure)
      if you try to use a mirror with an interleave of 0.

    * MIRRORing code enhanced:

	- When reading, CCD will issue the read from either side of the 
	  mirror.  CCD now implements a zone selector to try to optimize
	  which disk is selected based on seek proximity.

	- If a read error occurs, CCD will automatically try to recover by
	  reading from the other side of the mirror.  If this also fails, the
	  read error is propogated back.

    Note that CCD's mirroring has never been and still does not use any
    formal recovery mechanisms.  You cannot replace a disk in the mirror
    and expect to get something coherent out of it.  You can turn off a
    disk in the mirror by unmounting the CCD and reconfiguring it without
    the mirror disks as a non-mirror CCD, but that's about it.  Replacing
    a bad disk effectively requires backing up the good one, replacing
    the disk, and then regenerating (to be safe) both disks from the backup.

    The parity stuff never worked and probably never will.

						-Matt



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message