How to change a failed disk on a mirrored volume – Veritas Storage

Title in portuguese: “Como substituir um disco falho em um volume espelhado – Veritas Storage”.

In this article I will show how to change a failed disk on a Veritas Mirrored volume.

In this example we are assuming the VolM volume as a mirrored volume composed by two disks.

The volume is active and synch. We will simulate a physical failure in a disk that composes this mirrored volume.

We will replace this failed disk and synch the volume with a new disk.

To check that the volume is Active and Synch:

bash-3.00# vxprint
v  VolM         –            ENABLED  ACTIVE   20480    SELECT
pl VolM-01      VolM         ENABLED  20480    -        ACTIVE   –       -
sd c1t1d0-02    VolM-01      ENABLED  20480    0        -        -       -
pl VolM-02      VolM         ENABLED  ACTIVE   20480    CONCAT    -        RW
sd c1t2d0s2-02  VolM-02    ENABLED 10240    20480    0         c1t2d0   ENA

Let’s suppose that occurred a physical failure with one disk (c1t2d0).

To check the Volume Status:

bash-3.00# vxprint
v  VolM         fsgen        ENABLED  20480    -        ACTIVE   –       -
pl VolM-01      VolM         ENABLED  20480    -        ACTIVE   –       -
sd c1t1d0-02    VolM-01      ENABLED  20480    0        -        -       -
pl VolM-02      VolM         DISABLED 20480    -        NODEVICE –       -
sd c1t2d0s2-02  VolM-02      DISABLED 20480    0        NODEVICE –       -

At this point, you can notice that the volume is still active but the
Plex VolM-02 composed by a sub-disk from Disk c1t2d0 is not active due
to a physical problem. Volume VolM is still active because it was
configured using a mirror layout therefore the whole data is also stored
in the good disk (c1t1d0).
Note: If this volume was configured as a simple strippe volume, sorry, you would have no data now.

The next steps will describe how to add a new disk to Solaris in order to replace the failed one and recover the mirror.

1-)  Connect the new disk on the server

2-) Force solaris to recognize the new disk:

bash-3.00# devfsadm
bash-3.00# prtvtoc /dev/dsk/[device-name]

3-) To VxVM recognize the disk:

bash-3.00# vxdctl enable

The next step is to add this disk as the replacement of the failed disk.
Then, you need to add this new disk to a Disk Group using the same tag
as the failed disk.
In this example, faulted disk had the tag (c1t2d0s2).

4-) To add the new disk to the Disk Group DG1 as the replacement of the failed disk :

Note: -k force the system to take disk media from failed disk and assign to new disk.

bash-3.00# vxdg -k -g DG1 adddisk c1t2d0s2=c1t4d0

5-)To reattach the disk to the Disk Group:

bash-3.00# vxreattach

6-)To recover the volume and resynchronize the mirror:

bash-3.00# vxrecover -bs -g DG1 VolM

7-)To check if the volume is properly recovered:

bash-3.00# vxprint
v  VolM         fsgen        ENABLED  20480    -        ACTIVE   –       -
pl VolM-01      VolM         ENABLED  20480    -        ACTIVE   –       -
sd c1t1d0-02    VolM-01      ENABLED  20480    0        -        -       -
pl VolM-02      VolM         ENABLED  20480    -        ACTIVE   –       -
sd c1t2d0s2-02  VolM-02      ENABLED  20480    0        -        -

8-)To check that new disk is using the tag of failed disk:

bash-3.00# vxdisk -o alldgs list
DEVICE       TYPE            DISK         GROUP        STATUS
c1t0d0s2     auto:none       –            -            online invalid
c1t1d0s2     auto:cdsdisk    c1t1d0       DG1          online
c1t4d0s2     auto:cdsdisk    c1t2d0s2     DG1          online

Feel free to comment this article in order to improve its quality.

Reference:Veritas Storage Foundation 5.0 for UNIX Fundamentals.