Cant encapsulate disk

[zsowle1]

Hello All -

I have a Hitachi AMS 2300 array with 2tb presented to my solaris x64 host.   SF/HA 5.1, SP 1.   I am trying to encapsulate the device via vxdiskadm, and keep getting the following error.   I have searched, but cant find anything on this error.  Hope the experience in this forum can help me out.

The encapsulation operation failed with the following error:

  Arithmetic Exception - core dumped
  VxVM vxencap ERROR V-5-2-213
It is not possible to encapsulate ams_23000_1, for the following reason:
     <VxVM vxslicer ERROR V-5-1-5425 Cylinder size 0: Out of range>

Here is the vtoc of the device:

prtvtoc /dev/rdsk/c0t0d0
* /dev/rdsk/c0t0d0 partition map
*
* Dimensions:
*     512 bytes/sector
* 4294967295 sectors
* 4294967228 accessible sectors
*
* Flags:
*   1: unmountable
*  10: read-only
*
* Unallocated space:
*       First     Sector    Last
*       Sector     Count    Sector
*          34 4294950844 4294950877
*
*                          First     Sector    Last
* Partition  Tag  Flags    Sector     Count    Sector  Mount Directory
       2      5    00         34 2147483648 2147483681
       8     11    00  4294950878     16384 4294967261

Has anyone seen this before?  Any help would be huge.

[g_lee]

The prtvtoc of your disk appears to be incorrect, which could be the reason for the error being generated.

Slice 2 should span the whole disk (ie: start at sector 0 and be 4294967228 sectors long in the case of the disk above); however the output above shows it appears to start at sector 34, and is only 2147483648 sectors instead.

If the disk is not in use, try relabelling/repartitioning the disk to have the correct attributes for slice 2, then retry the encapsulation.

If the disk is in use, you may need to stop whatever process(es) are using the disk, and also take a backup before repartitioning the disk, in case this causes issues.

[zsowle1]

Howdy G-lee - thanks for the prompt response!

This is an x64 disk, with an EFI label.  (it will give slice 8 some sectors be default, and slice 2 will be reduced accordingly, and not start on cylinder 0).  I have changed the label to an SMI label via format -e, and used fdisk to carve out an 800gb partition.  I thought this would do the trick, but now im running into the following when I attempt to online the disk:

root@bells#vxdisk online ams_23000_1
VxVM vxdisk ERROR V-5-1-531 Device ams_23000_1: online failed:
        Device path not valid

Here is the vtoc - i am grasping at straws...  any suggestions would be huge!

root@bells#prtvtoc /dev/rdsk/c0t0d0s2
* /dev/rdsk/c0t0d0s2 partition map
*
* Dimensions:
*     512 bytes/sector
*     315 sectors/track
*     255 tracks/cylinder
*   80325 sectors/cylinder
*   21387 cylinders
*   21385 accessible cylinders
*
* Flags:
*   1: unmountable
*  10: read-only
*
* Unallocated space:
*       First     Sector    Last
*       Sector     Count    Sector
*       80325 1717669800 1717750124
*
*                          First     Sector    Last
* Partition  Tag  Flags    Sector     Count    Sector  Mount Directory
       2      5    00          0 1717750125 1717750124
       8      1    01          0     80325     80324

[g_lee]

Did you remove the disk before relabelling? The relabel from EFI to SMI would have modified the underlying devices/device paths so need to remove & rescan to pick up changes:

Try the following:

# vxdisk rm <disk>  #### ie: vxdisk rm ams_23000_1
# vxdisk scandisks
# vxdisk list
# vxdisk list <disk>

[zsowle1]

I read that the trial license doesnt activate DMP, and this could account for the "device path not valid" error.   I deactivated DMP via vxdiskadm.   So ams23000_1 has been changed to c0t0d0s2.

Thats a great call - I did remove the disks before re-adding, but keep getting the same error:

root@bells#vxdisk rm c0t0d0
root@bells#vxdisk list
DEVICE       TYPE            DISK         GROUP        STATUS
c2d0s2       auto:none       -            -            online invalid
c3d0s2       auto:none       -            -            online invalid
root@bells#vxdisk scandisks
root@bells#vxdisk list
DEVICE       TYPE            DISK         GROUP        STATUS
c0t0d0s2     auto            -            -            error
c2d0s2       auto:none       -            -            online invalid
c3d0s2       auto:none       -            -            online invalid
root@bells#vxdisk list c0t0d0s2
Device:    c0t0d0s2
devicetag: c0t0d0
type:      auto
flags:     error private autoconfig
pubpaths:  block=/dev/vx/dmp/c0t0d0s2 char=/dev/vx/rdmp/c0t0d0s2
guid:      -
udid:      -
site:      -
errno:     Device path not valid
Multipathing information:
numpaths:   1
c0t0d0s2        state=enabled

Once again, thanks for any suggestions you might have!

[zsowle1 · Accepted solution]

Through some trial and error, the final solution was to throw on an SMI disk label, and verify all cylinders were on slice 2.    Also, vxdiskadm didnt do the trick - I had to manually run vxdisksetup and then use vxdg to create the disk group.  Hope this helps someone out there!