>Number:         6676
>Category:       install
>Synopsis:       There is no rc support for RAIDframe
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    gnats-admin (GNATS administrator)
>State:          open
>Class:          change-request
>Submitter-Id:   net
>Arrival-Date:   Mon Dec 28 18:35:01 1998
>Last-Modified:
>Originator:     Eric S. Hvozda
>Organization:
	
>Release:        20Dec98 -current sources
>Environment:
	
System: NetBSD yale.apicom.com 1.3I NetBSD 1.3I (YALE) #0: Sun Dec 27 16:08:57 EST 1998 hvozda@yale.apicom.com:/usr/src/sys/arch/i386/compile/YALE i386


>Description:
	Now that we have RAIDframe, it would be nice if the system
	could be confiureed to use it at boot time automatically
	much like ccd(4).
>How-To-Repeat:
	Configure a RAIDframe device and create slices with file
	systems.  Come to the realization that there is no rc
	support for RAIDFrame.
>Fix:
	This patch does the job (I believe).  Currently there is
	no way to determine if the last shutdown was a clean one
	(I can think of a way, but it's a total hack dealing with
	fsck(8) which is probably not reasonable).

	One draw back of this method is that we always must
	rebuild parity.  This can be a big time sucker if the
	raid device in question is a *huge* one.  However,
	it's better than going forward with multi-user and
	unknown parity in RAIDframe:

*** etc/rc.dist	Mon Dec 28 15:03:33 1998
--- etc/rc	Mon Dec 28 20:58:07 1998
***************
*** 21,26 ****
--- 21,34 ----
  	ccdconfig -C
  fi
  
+ # Configure raid devices.
+ for dev in 0 1 2 3; do
+ 	if [ -f /etc/raid$dev.conf ]; then
+ 		raidctl -c /etc/raid$dev.conf raid$dev
+ 		raidctl -r raid$dev
+ 	fi
+ done
+ 
  # Add all block-type swap devices; these might be necessary
  # during disk checks.
  swapctl -A -t blk
>Audit-Trail:
>Unformatted: