Solstice Disksuite FAQ(10)

diandianrong

Document ID: 2085
SYNOPSIS:  Ten common Solstice Disksuite questions
DETAIL DESCRIPTION:
1.  DiskSuite versions and compatibility
2.  How to populate md.conf file
3.  How to suppress forceload message
4.  Replicas on SSA marked as 'bad'
5.  Checking metadevice states
6.  Failed boot under DiskSuite control
7.  Checking replica states
8.  Explanation of disksets
9.  Increase the number of metadevices
10.  Metainit command for RAID5
SOLUTION SUMMARY:
1.  Which version of DiskSuite runs on SunOS 4.1.4?
    Online: DiskSuite 1.0 is the only version of DiskSuite compatible
    with SunOS 4.X.
        
    While installing DiskSuite 1.0, you may encounter an error:
    This software is not compatible with the current operating system.
    This is the list of compatible operating systems:
                  4.1.1
                  4.1.2
    Do you want to continue? [y|n]
    Simply answer "y" and continue with the installation.
    DiskSuite 1.0 is compatible with SunOS 4.1.2, 4.1.2, 4.1.3, and 4.1.4.
    Online DiskSuite 2.0, 2.0.1 and 3.0 are compatible with Solaris 2.1
    and later releases of Solaris 2.X.
    Solstice Disksuite 4.0 is compatible with Solaris 2.3 and later
    Solaris 2.X releases.
    Solstice Disksuite 4.1 is compatible with Solaris 2.4 and later
    Solaris 2.X releases.
2.  After setting up Metadevices using metatool, how do I record
    the configuration in the md.tab file?"
    Run the following command after the configuration is complete
    to record the configuration:
  
    metastat -p > /etc/opt/SUNWmd/md.tab
3.  During the boot process, the message "WARNING: forceload of
    misc/md_hotspares failed." appears, what does it mean?
                                         
    This warning is harmless and can be ignored.  The forceload statement
    in the /etc/system file is designed to load the device driver as
    soon as possible in the boot process.  If the device driver is not
    needed, it will not load, and reports this type of warning message.
    To suppress the message, create an empty hotspare pool with the command:
    # metainit hsp001
4.  Upon reboot, all replicas on the SPARCStorage Array go bad.
    How do I prevent this?
        
    Add the following entries to the /etc/system file to avoid this problem:
     forceload: drv/ssd
     forceload: drv/SUNW,pln
     forceload: drv/SUNW,soc
    You add forceloads to the /etc/system file so that the SSA 200 disks are
    available early in the boot process.
        
5.  A disk went bad and needs to be replaced, how do I find the
    state of the DiskSuite software?
    Use the metastat command to find the state of the Disksuite configuration.
    If a component is in need of maintenance, the metastat output will
    give the course of action required.
6.  How do I remove the boot disk from DiskSuite control?
        ok boot cdrom
        # fsck   /dev/rdsk/c0....
        # mount /dev/dsk/c0t3d0s0 /a
        Note: this assumes that your boot device is target three on
        controller zero.
        # vi /a/etc/system
            comment

out entries pertaining to / being a metadevice
            (rootdev:pseudo/md:0,blk)
        # vi /a/etc/vfstab
            change metadevice name to physical disk partition for all
            file systems associated with booting
        # mv /a/etc/opt/SUNWmd/etc/mddb.cf /a/etc/opt/SUNWmd/prob.mddb.cf
        # umount /a
        # fsck /dev/rdsk/c0t3d0s0
        # reboot
7.  How do I check the status of the state database replicas?
    Use the metadb command to find the state of the Disksuite replicas.
    If a replica is in need of attention, the flags will report appropriately.
    Often the SDS metatool will report the replicas to be in a critical
    state when metadb reports no problems.
        # metadb -i
                flags           first blk       block count
             a m  pc luo        16              1034 /dev/dsk/c1t5d0s4
             a    pc luo        1050            1034   /dev/dsk/c1t5d0s4
             a    pc luo        2084            1034  /dev/dsk/c1t5d0s4
         o - replica active prior to last mddb configuration change
         u - replica is up to date
         l - locator for this replica was read successfully
         c - replica's location was in /etc/opt/SUNWmd/mddb.cf
         p - replica's location was patched in kernel
         m - replica is master, this is replica selected as input
         W - replica has device write errors
         a - replica is active, commits are occurring to this replica
         M - replica had problem with master blocks
         D - replica had problem with data blocks
         F - replica had format problems
         S - replica is too small to hold current data base
         R - replica had device read errors
8. How are disksets used?
   A shared diskset is a grouping of two hosts and disk drives in which all
   the drives are accessible by both hosts.  Disksuite requires that the
   device name be identical on each host in the diskset.  There is one
   metadevice state database per shared diskset and one on the local disks
   of each host.
   Before you can begin to create or administer disksets, the disksuite
   software must be installed on each host in the diskset and each host
   must have local database replicas set up.  In addition, all disks that
   you plan to share between hosts in the diskset must be connected to each
   host and must have the same name (shared_diskset name) on each host.
9. How do I create more than 128 meta devices?
   Edit the file /kernel/drv/md.conf file and change the "nmd" field from
   nmd=128 to whatever amount is needed (i.e., 255).
        
   Then boot -r to build these new devices.
10. How do I metainit a RAID5 device?
    metainit d0 -r c2t0d0s4 c2t2d0s4 c2t4d0s4 c2t0d4s4 c2t3d1s4 -k -i 32b
    If you omit the -k option data loss will occur.
    This only applies to existing RAID5 metadevices, not new ones.
DATE APPROVED: 03/25/98
KEYWORDS:      disksuite forceload RAID5
OS RELEASE:    N/A


本文来自ChinaUnix博客，如果查看原文请点：http://blog.chinaunix.net/u/26090/showart_321339.html