Return to set-up_raidframe.mdwn CVS log

Up to [NetBSD Developer Wiki] / wikisrc

Xref the man-pages.

    1: [[!meta  title="Setting up raidframe(4) on NetBSD"]]
    2: 
    3: After posting lots of questions, and getting lots of expert help, I finally got my new server up and running with raid!  Since I asked so many questions, it was suggested that I recap the process so that someone else might be able to make it work without having to ask!
    4: 
    5: So ...
    6: 
    7: Here's what I did recently to get  [[!template id=man name="raid" section="4"]] up and running on my new server.
    8: 
    9: 1. This is really important!  Read the man page for  [[!template id=man name="raidctl" section="8"]].  It is a long man page, but it is important to understand how things work.  And there are some actual examples towards the end, which will only make sense if you've read the intro material.  Don't be intimidated by all the detail, just RTFM before you start!
   10: 
   11: 2. Use [[!template id=man name="dd" section="1"]] to completely erase the drives.  This is useful to make sure that any existing labels/wedges/etc. get removed, and also makes it easier to initialize the raidset parity.  On Western Digital SATA-II 500GB drive, it took less than two hours (each) for this step.  (The operation starts off at ~110GB/sec, but will get slower as the heads get closer to the center of the drive where there are fewer sectors per cylinder.)
   12: 
   13:     	dd if=/dev/zero bs=32k of=/dev/rwd2a
   14:     	dd if=/dev/zero bs=32k of=/dev/rwd3a
   15: 
   16: 3. This step might not be necessary but I wasn't taking any chances!  Also, this step is only applicable on ports which use DOS disk layouts.  For other ports, you should be able to skip this step, and proceed directly to Step 4.
   17: 
   18:  Use  [[!template id=man name="fdisk" section="8"]] to set up DOS/Windows partition data.  Make sure you use the "-A 2048" option for alignment.
   19: 
   20:     	fdisk -uai -A 2048 wd2a
   21:     	fdisk -uai -A 2048 wd3a
   22: 
   23:  (If you are using an older version of fdisk and get the error
   24: 
   25:     	fdisk: Bad argument to the -A flag.
   26: 
   27:  you can specify "-A 2048/2048" instead.)
   28: 
   29:  Set the partition 0 type to 169 (NetBSD), and let it occupy the entire drive.  If you're going to boot from this raidset, you should also make this the active partition, and update the boot blocks when prompted.
   30: 
   31:  When finished, this is what my drive reports:
   32: 
   33:     	Partitions aligned to 16065 sector boundaries, offset 63
   34: 
   35:     	Partition table:
   36:     	0: NetBSD (sysid 169)
   37:     	    start 2048, size 976771120 (476939 MB, Cyls 0/32/33-60801/80/63),
   38:     		Active
   39:     	        PBR is not bootable: All bytes are identical (0x00)
   40:     	1: <UNUSED>
   41:     	2: <UNUSED>
   42:     	3: <UNUSED>
   43:     	Bootselector disabled.
   44:     	First active partition: 0
   45: 
   46:  (The "PBR is not bootable" line will not appear if you later use [[!template id=man name="installboot" section="8"]] - see step 14 below.)
   47: 
   48: 4. Use  [[!template id=man name="disklabel" section="8"]] to provide a NetBSD disklabel on each member of the raidset.  Create an 'e' partition that describes the whole disk - so it is essentially a copy of the 'c' partition.  Make sure that the offset of your 'e' partition is large enough to accomodate any architecture-specific bootstrap requirements, and that the offset is a multiple of the hard drive's native transfer size.
   49: 
   50:  Set the fstype on your 'e' partition to RAID (4.2BSD will also work, but not if you want to use RAID_AUTOCONFIG).
   51: 
   52:  Here's my partition info:
   53: 
   54:     	5 partitions:
   55:     	#        size  offset  fstype [fsize bsize cpg/sgs]
   56:     	 c: 976771120    2048  unused      0     0        # (Cyl.    2*- 969020)
   57:     	 d: 976773168       0  unused      0     0        # (Cyl.    0 - 969020)
   58:     	 e: 976771120    2048    RAID                     # (Cyl.    2*- 969020)
   59: 
   60: 5. Now create a raid.conf file to describe the raidset.  Mine is a simple RAID-1 (mirror):
   61: 
   62:     	START array
   63:     	#numrow numcol numspare
   64:     	1 2 0
   65: 
   66:     	# Identify physical disks
   67:     	START disks
   68:     	/dev/wd2e
   69:     	/dev/wd3e
   70: 
   71:     	# Layout is simple - 64 sectors per stripe
   72:     	START layout
   73:     	#Sect/StripeUnit StripeUnit/ParityUnit StripeUnit/ReconUnit RaidLevel
   74:     	128 1 1 1
   75: 
   76:     	# No spares
   77:     	#START spare
   78: 
   79:     	# Command queueing
   80:     	START queue
   81:     	fifo 100
   82: 
   83: 6. Create the raidset using  [[!template id=man name="raidctl" section="8"]].  Until you do this for the first time, there are no raid "component labels" on the member drives, so you will need to use the "-C" option;  if the raidset has been previously created, you should use "-c" instead (this is what /etc/rc.d/raidframe uses at system startup).
   84: 
   85:     	raidctl -C /etc/raid1.conf raid1
   86: 
   87: 7. Next, initialize the component labels with raidctl.
   88: 
   89:     	raidctl -I 201105061405 raid1
   90: 
   91:  (For serial number, I just used the current date/time; any 64-bit integer will work.)
   92: 
   93: 8. Now use raidctl to initialize the parity of the raidset.  On my 500GB drives, this took about 1.5 hours.
   94: 
   95:     	raidctl -i raid1
   96: 
   97:  It won't tie up your session, as the parity operation happens in the kernel in the background.  You can monitor the progress using
   98: 
   99:     	raidctl -s raid1
  100: 
  101: 9. When parity is finished, you can treat the raidset as a normal disk. You will need to write a NetBSD disklabel (this label is separate from the labels on the physical drives).  Make certain that all the offsets are multiples of the drive's native block size; otherwise you will have unaligned transfers and performance will be very bad!  (This is also why, in step 3 above, we use "-A 2048" for fdisk rather than letting it default to 63!")
  102: 
  103:  Here is what my drive looks like:
  104: 
  105:     	7 partitions:
  106:     	#        size    offset  fstype [fsize bsize cpg/sgs]
  107:     	 c: 976770944         0  unused      0     0       # (Cyl.      0 - 953877*)
  108:     	 d: 976770944         0  unused      0     0       # (Cyl.      0 - 953877*)
  109:     	 e: 536870912         0  4.2BSD   2048 16384    0  # (Cyl.      0 - 524287)
  110:     	 f: 104857600 536870912  4.2BSD   2048 16384    0  # (Cyl. 524288 - 626687)
  111:     	 g: 335042432 641728512  4.2BSD   2048 16384    0  # (Cyl. 626688 - 953877*)
  112: 
  113: 10. Now, just as you would for a real disk, use newfs to initialize the filesystems.  Then mount them and start using your new raidset!
  114: 
  115:     	newfs /dev/raid1e
  116:     	mount /dev/raid1e /mnt
  117: 
  118: 11. If your kernel includes the RAID_AUTOCONFIG option, you can use raidctl to let your system automatically reconfigure the raidset at every boot.  (If you don't do this, system startup will still take care of this, as long as your configuration file is properly named "raid[0-9].conf" or "raid[1-9][0-9].conf" and is located in /etc/ directory.)
  119: 
  120:     	raidctl -A yes raid1
  121: 
  122: 12. If you're going to boot from the raidset, you probably want to have an 'a' partition.  My bootable raidset has members wd0 and wd1, which are labeled like this:
  123: 
  124:     	5 partitions:
  125:     	#        size    offset     fstype [fsize bsize cpg/sgs]
  126:     	 c: 488395120      2048     unused      0     0        # (Cyl.      2*- 484520)
  127:     	 d: 488397168         0     unused      0     0        # (Cyl.      0 - 484520)
  128:     	 e: 488395120      2048       RAID                     # (Cyl.      2*- 484520)
  129: 
  130: 13. And the raidset itself has these partitions:
  131: 
  132:     	6 partitions:
  133:     	#        size    offset  fstype [fsize bsize cpg/sgs]
  134:     	 a:  41943040         0  4.2BSD   2048 16384    0  # (Cyl.      0 -  40959)
  135:     	 b:  62914560  41943040    swap                    # (Cyl.  40960 - 102399)
  136:     	 c: 488395008         0  unused      0     0       # (Cyl.      0 - 476948*)
  137:     	 d: 488395008         0  unused      0     0       # (Cyl.      0 - 476948*)
  138:     	 e: 125829120 104857600  4.2BSD   2048 16384    0  # (Cyl. 102400 - 225279)
  139:     	 f: 257708288 230686720  4.2BSD   2048 16384    0  # (Cyl. 225280 - 476948*)
  140: 
  141: 14. You will also need to run  [[!template id=man name="installboot" section="8"]] to make the raidset bootable. Do this for each member drive.  Please note that installboot is run for the RAID partition on each member disk, and NOT on the raidn disk!  This also assumes that the RAID partition (the 'e' partition, see step 4 above) starts at the beginning of the whole-NetBSD 'c' partition.  If you don't do this, I don't think that the i386 MBR boot code will be able to find your bootstrap image.
  142: 
  143:     	mount /dev/raid0a /mnt
  144:     	installboot /dev/wd0e /usr/mdec/bootxx_ffsv1 /mnt/boot
  145:     	installboot /dev/wd1e /usr/mdec/bootxx_ffsv1 /mnt/boot
  146: 
  147: 15. Finally, you need to make sure your kernel has the RAID_AUTOCONFIG option, and enable the raidset as the root device:
  148: 
  149:     	raidctl -A root raid0