test

Sun Cluster 2.2 System Administration Guide

Replacing a SPARCstorage Array Disk

This section describes replacing a SPARCstorage Array (SSA) multihost disk without interrupting Sun Cluster services (online replacement) when the volume manager is reporting problems such as:

How to Replace a SPARCstorage Array Disk (Solstice DiskSuite)

These are the high-level steps to replace a multihost disk in a Solstice DiskSuite configuration. Some of the steps in this procedure apply only to configurations using SPARCstorage Array 100 series or SPARCstorage Array 200 series with the differential SCSI tray.

These are the detailed steps to replace a failed multihost disk in a Solstice DiskSuite configuration.

  1. Switch ownership of the affected logical hosts to other nodes by using the haswitch(1M) command.


    phys-hahost1# haswitch phys-hahost1 hahost1 hahost2

    The SPARCstorage Array tray containing the failed disk might contain disks included in more than one logical host. If this is the case, switch ownership of all logical hosts with disks using this tray to another node in the cluster.

  2. Identify the disk to be replaced by examining metastat(1M) and /var/adm/messages output.

    When metastat(1M) reports that a device is in maintenance state or some of the components have been replaced by hot spares, you must locate and replace the device. A sample metastat(1M) output follows. In this example, device c3t3d4s0 is in maintenance state.


    phys-hahost1# metastat -s hahost1
...
 d50:Submirror of hahost1/d40
      State: Needs Maintenance
      Stripe 0:
          Device       Start Block      Dbase      State          Hot Spare
          c3t3d4s0     0                No         Okay           c3t5d4s0
...

    Check /var/adm/messages to see what kind of problem has been detected.


    ...
Jun 1 16:15:26 host1 unix: WARNING: /io-unit@f,e1200000/sbi@0.0/SUNW,pln@a0000000,741022/ssd@3,4(ssd49):   
Jun 1 16:15:26 host1 unix: Error for command `write(I))' Err
Jun 1 16:15:27 host1 unix: or Level: Fatal
Jun 1 16:15:27 host1 unix: Requested Block 144004, Error Block: 715559
Jun 1 16:15:27 host1 unix: Sense Key: Media Error
Jun 1 16:15:27 host1 unix: Vendor `CONNER':
Jun 1 16:15:27 host1 unix: ASC=0x10(ID CRC or ECC error),ASCQ=0x0,FRU=0x15
...

  3. Determine the location of the problem disk by running the luxadm(1M) command.

    The luxadm(1M) command lists the trays and the drives associated with them. The output differs for each SPARCstorage Array series. This example shows output from a SPARCstorage Array 100 series array. The damaged drive is highlighted below. 


    phys-hahost1# luxadm display c3
         SPARCstorage Array Configuration
Controller path:

/devices/iommu@f,e0000000/sbus@f,e0001000/SUNW,soc@0,0/SUNW,pln@a0000000,779a16:ctlr
         DEVICE STATUS
         TRAY1          TRAY2          TRAY3
Slot
1        Drive:0,0      Drive:2,0      Drive:4,0
2        Drive:0,1      Drive:2,1      Drive:4,1
3        Drive:0,2      Drive:2,2      Drive:4,2
4        Drive:0,3      Drive:2,3      Drive:4,3
5        Drive:0,4      Drive:2,4      Drive:4,4
6        Drive:1,0      Drive:3,0      Drive:5,0
7        Drive:1,1      Drive:3,1      Drive:5,1
8        Drive:1,2      Drive:3,2      Drive:5,2
9        Drive:1,3      Drive:3,3      Drive:5,3
10       Drive:1,4      Drive:3,4      Drive:5,4

         CONTROLLER STATUS
Vendor:    SUN
Product ID:  SSA110
Product Rev: 1.0
Firmware Rev: 3.9
Serial Num: 000000741022
Accumulate performance Statistics: Enabled

  4. Detach all submirrors with components on the disk being replaced.

    If you are detaching a submirror that has a failed component, you must force the detach using the metadetach -f command. The following example command detaches submirror d50 from metamirror d40.


    phys-hahost1# metadetach -s hahost1 -f d40 d50

  5. Use the metaclear(1M) command to clear the submirrors detached in Step 4.


    phys-hahost1# metaclear -s hahost1 -f d50

  6. Before deleting replicas and hot spares, make a record of the location (slice), number of replicas, and hot spare information (names of the devices and list of devices that contain hot spare pools) so that the actions can be reversed following the disk replacement.

  7. Delete all hot spares that have Available status and are in the same tray as the problem disk.

    This includes all hot spares, regardless of their logical host assignment. In the following example, the metahs(1M) command reports hot spares on hahost1, but shows that none are present on hahost2" 


    phys-hahost1# metahs -s hahost1 -i
hahost1:hsp000 2 hot spares
        c1t4d0s0                Available       2026080 blocks
        c3t2d5s0                Available       2026080 blocks
phys-hahost1# metahs -s hahost1 -d hsp000 c3t2d4s0 
hahost1:hsp000:
        Hotspare is deleted
phys-hahost1# metahs -s hahost2 -i
phys-hahost1#
hahost1:hsp000 1 hot spare
			c3t2d5s0                Available       2026080 blocks

  8. Use the metaset(1M) command to remove the failed disk from the diskset.

    The syntax for the command is shown below. In this example, diskset is the name of the diskset containing the failed disk, and drive is the DID name of the disk in the form dN (for new installations of Sun Cluster), or cNtYdZ (for installations that upgraded from HA 1.3).


    # metaset -s diskset -d drive

    This operation can take up to fifteen minutes or more, depending on the size of your configuration and the number of disks.

  9. Delete any metadevice state database replicas that are on disks in the tray to be serviced.

    The metadb(1M) command with the -s option reports replicas in a specified diskset. 


    phys-hahost1# metadb -s hahost1
phys-hahost1# metadb -s hahost2
phys-hahost1# metadb -s hahost1 -d replicas-in-tray
phys-hahost1# metadb -s hahost2 -d replicas-in-tray

  10. Locate the submirrors using components that reside in the affected tray.

    One method is to use the metastat(1M) command to create temporary files that contain the names of all metadevices. For example:


    phys-hahost1# metastat -s hahost1 > /usr/tmp/hahost1.stat
phys-hahost1# metastat -s hahost2 > /usr/tmp/hahost2.stat

    Search the temporary files for the components in question (c3t3dn and c3t2dn in this example). The information in the temporary files will look like this:


    ...
hahost1/d35: Submirror of hahost1/d15
   State: Okay
   Hot Spare pool: hahost1/hsp100
   Size: 2026080 blocks
   Stripe 0:
      Device      Start Block     Dbase     State      Hot Spare
      c3t3d3s0    0               No        Okay       
hahost1/d54: Submirror of hahost1/d24
   State: Okay
   Hot Spare pool: hahost1/hsp106
   Size: 21168 blocks
   Stripe 0:
      Device      Start Block     Dbase     State      Hot Spare
      c3t3d3s6    0               No        Okay       
...

  11. Take offline all other submirrors that have components in the affected tray.

    Using the output from the temporary files in Step 10, run the metaoffline(1M) command on all submirrors in the affected tray


    phys-hahost1# metaoffline -s hahost1 d15 d35
phys-hahost1# metaoffline -s hahost1 d24 d54
...
    .

    Run metaoffline(1M) as many times as necessary to take all the submirrors off line. This forces Solstice DiskSuite to stop using the submirror components.

  12. If enabled, flush the NVRAM on the controller, tray, individual disk or disks.


    phys-hahost1# luxadm sync_cache pathname

    A confirmation appears, indicating that NVRAM has been flushed. See "Flushing and Purging NVRAM", for details on flushing NVRAM data.

  13. Spin down all disks in the affected SPARCstorage Array tray(s).

    Use the luxadm stop command to spin down the disks. Refer to the luxadm(1M) man page for details.


    phys-hahost1# luxadm stop -t 2 c3

    Caution - Caution -

    Do not run any Solstice DiskSuite commands while a SPARCstorage Array tray is spun down because the commands might have the side effect of spinning up some or all of the drives in the tray.

  14. Replace the disk.

    Refer to the hardware service manuals for your SPARCstorage Array for details on this procedure.

  15. Update the DID driver's database with the new device ID.

    Use the -l flag to scdidadm(1M) to identify the DID name for the lower-level device name of the drive to be replaced. Then update the DID drive database using the -R flag to scdidadm(1M). Refer to the Sun Cluster 2.2 Software Installation Guide for details on the DID pseudo driver.


    phys-hahost1# scdidadm -o name -l /dev/rdsk/c3t3d4
6	phys-hahost1:/dev/rdsk/c3t3d4	/dev/did/rdsk/d6
phys-hahost1# scdidadm -R d6

  16. Make sure all disks in the affected multihost disk expansion unit spin up.

    The disks in the multihost disk expansion unit should spin up automatically. If the tray fails to spin up within two minutes, force the action by using the following command:


    phys-hahost1# luxadm start -t 2 c3

  17. Add the new disk back into the diskset by using the metaset(1M) command.

    This step automatically adds back the number of replicas that were deleted from the failed disk. The command syntax is as follows, where diskset is the name of the diskset containing the failed disk, and drive is the DID name of the disk in the form dN (for new installations of Sun Cluster), or cNtYdZ (for installations that upgraded from HA 1.3):


    # metaset -s diskset -a drive

  18. (Optional) If you deleted replicas that belonged to other disksets from disks that were in the same tray as the errored disk, use the metadb(1M) command to add back the replicas.


    phys-hahost1# metadb -s hahost2 -a deleted-replicas

    To add multiple replicas to the same slice, use the -c option.

  19. Use the scadmin(1M) command to reserve and enable failfast on the specified disk that has just been added to the diskset.


    phys-hahost2# scadmin reserve c3t3d4

  20. Use the format(1M) or fmthard(1M) command to repartition the new disk.

    Make sure that you partition the new disk exactly as the disk that was replaced. (Saving the disk format information was recommended in Chapter 1, Preparing for Sun Cluster Administration.)

  21. Use the metainit(1M) command to reinitialize disks that were cleared in Step 5. 


    phys-hahost1# metainit -s hahost1 d50

  22. Bring online all submirrors that were taken off line in Step 11.


    phys-hahost1# metaonline -s hahost1 d15 d35
phys-hahost1# metaonline -s hahost1 d24 d54
...

    Run the metaonline(1M) command as many times as necessary to bring online all the submirrors.

    When the submirrors are brought back online, Solstice DiskSuite automatically performs resyncs on all the submirrors, bringing all data up-to-date.

    Note -

    Running the metastat(1M) command at this time would show that all metadevices with components residing in the affected tray are resyncing.

  23. Attach submirrors that were detached in Step 4.

    Use the metattach(1M) command to perform this step. See the metattach(1M) man page for details. 


    phys-hahost1# metattach -s hahost1 d40 d50

  24. Replace any hot spares in use in the submirrors attached in Step 23.

    If a submirror had a hot spare replacement in use before you detached the submirror, this hot spare replacement will be in effect after the submirror is reattached. This step returns the hot spare to Available status.


    phys-hahost1# metareplace -s hahost1 -e d40 c3t3d4s0

  25. Restore all hot spares that were deleted in Step 7.

    Use the metahs(1M) command to add back the hot spares. See the metahs(1M) man page for details.


    phys-hahost1# metahs -s hahost1 -a hsp000 c3t2d5s0

  26. If necessary, switch logical hosts back to their default masters.


    phys-hahost1# haswitch phys-hahost2 hahost2

  27. Verify that the replacement corrected the problem.


    phys-hahost1# metastat -s hahost1

How to Replace a SPARCstorage Array Disk (VxVM)

In a VxVM configuration, it is possible to replace a SPARCstorage Array disk without halting the system, as long as the configuration is mirrored.

Note -

If you need to replace a disk in a bootable SPARCstorage Array, do not remove the SSA trays containing the boot disk of the hosts. Instead, shut down the host whose boot disk is present on that tray. Let the cluster software reconfigure the surviving nodes for failover to take effect before servicing the faulty disk. Refer to the SPARCstorage Array User's Guide for more information.

These are the high-level steps to replace a multihost disk in an VxVM environment using SPARCstorage Array 100 series disks.

These are the detailed steps to replace a multihost disk in an VxVM environment using SPARCstorage Array 100 series disks.

  1. If the replaced disk is a quorum device, use the scconf -q command to change the quorum device to a different disk.

  2. Identify all the volumes and corresponding plexes on the disks in the tray which contains the faulty disk.

    1. From the physical device address cNtNdN, obtain the controller number and the target number.

      For example, if the device address is c3t2d0, the controller number is 3 and the target is 2.

    2. Identify devices from a vxdisk list output.

      If the target is 0 or 1, identify all devices with physical addresses beginning with cNt0 and cNt1, where N is the controller number. If the target is 2 or 3, identify all devices with physical addresses beginning with cNt2 and cNt3. If the target is 4 or 5, identify all devices with physical addresses beginning with cNt4 and cNt5. Here is an example of how vxdisk can be used to obtain the information.


      # vxdisk -g diskgroup -q list | egrep c3t2\|c3t3 | nawk '{print $3}'

    3. Record the volume media name for the faulty disk from the output of the command.

      You will need this name in Step 10.

    4. Identify all plexes on the above devices by using the appropriate version (csh, ksh, or Bourne shell) of the following command.


      PLLIST=`vxprint -ptq -g diskgroup -e '(aslist.sd_dm_name in 

("c3t2d0","c3t3d0","c3t3d1")) && (pl_kstate=ENABLED)' | nawk '{print $2}'`

      For csh, the syntax is set PLLIST .... For ksh, the syntax is export PLLIST= .... The Bourne shell requires the command export PLLIST after the variable is set.

  3. After you have set the variable, stop I/O to the volumes whose components (subdisks) are on the tray.

    Make sure all volumes associated with that tray are detached (mirrored or RAID5 configurations) or stopped (simple plexes). Issue the following command to detach a mirrored plex.


    # vxplex det ${PLLIST}

    An alternate command for detaching each plex in a tray is:


    # vxplex -g diskgroup -v volume det plex

    To stop I/O to simple plexes, unmount any file systems or stop database access.

    Note -

    Mirrored volumes will still be active because the other half of the mirror is still available.

  4. Remove the disk from the disk group.


    # vxdg -g diskgroup rmdisk diskname

  5. Spin down the disks in the tray.


    # luxadm stop -t tray controller

  6. Replace the faulty disk.

  7. Spin up the drives.


    # luxadm start -t tray controller

  8. Initialize the replacement disk.


    # vxdisksetup -i devicename

  9. Scan the current disk configuration again.

    Enter the following commands on all nodes in the cluster.


    # vxdctl enable
# vxdisk -a online

  10. Add the new disk to the disk group.

    The device-media-name is the volume media name recorded in Step 2c.


    # vxdg -g diskgroup -k adddisk device-media-name=device-name

  11. Resynchronize the volumes.


    # vxrecover -g diskgroup -b -o