Replacing a failed disk with DiskSuite

Example one

In this example, the boot disk mirror c0t1d0 failed. All submirrors on c0t1d0 were placed in "maintenance" state, so no reads or writes were occurring on the disk. The replacement disk has identical geometry of boot disk c0t0d0.

1. Delete any state database replicas from the failed disk. A "W" in metadb output indicates replica device write errors.

# metadb
        flags           first blk       block count
     a m  p  luo        16              8192            /dev/dsk/c0t0d0s6
     a    p  luo        8208            8192            /dev/dsk/c0t0d0s6
      W   p  l          16              8192            /dev/dsk/c0t1d0s6
      W   p  l          8208            8192            /dev/dsk/c0t1d0s6

# metadb -d /dev/dsk/c0t1d0s6
# metadb
        flags           first blk       block count
     a m  p  luo        16              8192            /dev/dsk/c0t0d0s6
     a    p  luo        8208            8192            /dev/dsk/c0t0d0s6

2. Replace the failed disk.

3. Copy the partition table from the good disk to the replacement disk.
# prtvtoc /dev/rdsk/c0t0d0s2 | fmthard -s - /dev/rdsk/c0t1d0s2

4. Create state database replicas on the replacement disk.
# metadb -a -c 2 /dev/rdsk/c0t1d0s6

5. Determine which submirrors need to be resynchronized.

# metastat | grep 'Invoke: metareplace'
    Invoke: metareplace d30 c0t1d0s0 <new device>
    Invoke: metareplace d31 c0t1d0s1 <new device>
    Invoke: metareplace d33 c0t1d0s3 <new device>
    Invoke: metareplace d34 c0t1d0s4 <new device>
    Invoke: metareplace d35 c0t1d0s5 <new device>

6. Resynchronize the submirrors.

# ./metareplace -e d30 c0t1d0s0
d30: device c0t1d0s0 is enabled
# ./metareplace -e d31 c0t1d0s1
d31: device c0t1d0s1 is enabled
# ./metareplace -e d33 c0t1d0s3
d32: device c0t1d0s3 is enabled
# ./metareplace -e d34 c0t1d0s4
d33: device c0t1d0s4 is enabled
# ./metareplace -e d35 c0t1d0s5
d34: device c0t1d0s5 is enabled

You can monitor the resynchronization progress with metastat.

Example two

In this example, two of the four slices on c1t1d0 have been placed in "maintenance" state. Because of the large number of transport errors, I chose to replace the disk. Since two of the slices are in "okay" state--reads and writes are occurring on the slices--they have to be taken offline before replacing the disk.

# iostat -En
...
c1t1d0          Soft Errors: 1 Hard Errors: 1 Transport Errors: 28
Vendor: SEAGATE  Product: ST373405FSUN72G  Revision: 0638 Serial No: 0202K0ZZ0F
Size: 73.40GB <73400057856 bytes>
Media Error: 0 Device Not Ready: 0 No Device: 1 Recoverable: 1

# metastat d20 d21 d23 d25
d20: Concat/Stripe
    Size: 3154560 blocks
    Stripe 0:
        Device              Start Block  Dbase State        Hot Spare
        c1t1d0s0                   0     No    Maintenance 

d21: Concat/Stripe
    Size: 8395200 blocks
    Stripe 0:
        Device              Start Block  Dbase State        Hot Spare
        c1t1d0s1                   0     No    Okay        

d23: Concat/Stripe
    Size: 2106432 blocks
    Stripe 0:
        Device              Start Block  Dbase State        Hot Spare
        c1t1d0s3                   0     No    Maintenance 

d25: Concat/Stripe
    Size: 129672768 blocks
    Stripe 0:
        Device              Start Block  Dbase State        Hot Spare
        c1t1d0s5                   0     No    Okay        

Before replacing the disk, run metaoffline on the submirrors that are in "okay" state.

# metaoffline d1 d21
d1: submirror d21 is offlined
# metaoffline d5 d25
d5: submirror d25 is offlined

Delete the state database replaces on c1t1d0.

# metadb
        flags           first blk       block count
     a m  p  luo        16              1034            /dev/dsk/c1t0d0s4
     a    p  luo        16              1034            /dev/dsk/c1t0d0s6
     a    p  luo        16              1034            /dev/dsk/c1t1d0s4
     a    p  luo        16              1034            /dev/dsk/c1t1d0s6
# metadb -d /dev/dsk/c1t1d0s4 /dev/dsk/c1t1d0s6

# metadb
        flags           first blk       block count
     a m  p  luo        16              1034            /dev/dsk/c1t0d0s4
     a    p  luo        16              1034            /dev/dsk/c1t0d0s6

Replace the disk with a disk of identical geometry, and copy the partition table from the good disk to the replacement disk.

# prtvtoc /dev/rdsk/c1t0d0s2 | fmthard -s - /dev/rdsk/c1t1d0s2
fmthard:  New volume table of contents now in place.

Re-create the state database replicas on the replacement disk.

# metadb -a -c 1 /dev/dsk/c1t1d0s4 /dev/dsk/c1t1d0s6

Run metareplace to replace the failed submirrors that were in "maintenance" state and metaonline to re-enable the submirrors that were in "okay" state.

# metastat | grep metareplace
    Invoke: metareplace d0 c1t1d0s0 <new device>
    Invoke: metareplace d3 c1t1d0s3 <new device>
# metareplace -e d0 c1t1d0s0
d0: device c1t1d0s0 is replaced with c1t1d0s0
# metareplace -e d3 c1t1d0s3
d3: device c1t1d0s3 is replaced with c1t1d0s3

# metaonline d1 d21
d1: submirror d21 is onlined
# metaonline d5 d25
d5: submirror d25 is onlined

Monitor the output of metastat as the submirrors resynchronize. If the submirrors re-enabled with metaonline are placed in "maintenance" state, use metareplace on these submirrors.

# metareplace -e d5 c1t1d0s5
d5: device c1t1d0s5 is replaced with c1t1d0s5

Back to brandonhutchinson.com.
Last modified: 2007/03/20