Error Message

The operating system panic() stack trace for this issue is:

(divide by zero)  rp: 0x2a106cba500  --
  addr: 0x7bf3c808
pc:  0x7bf3c808 vxio:vol_splitsio_iogen+0xd4:   divx	%i3, %i1, %l1
npc: 0x7bf3c80c vxio:vol_splitsio_iogen+0xd8:   stx	%l1, [%l7 + 0x130]
  global:			%g1               0x10
	%g2                  0	%g3           0x37856f
	%g4                  0	%g5          0xc2422e0
	%g6                  0	%g7      0x3003415a160
  out:	%o0      0x60086a0d060	%o1          0x19bdb28
	%o2                0x4	%o3                  0
	%o4            0x10000	%o5          0x1000000
	%sp      0x2a106cb9da1	%o7         0x7bf3c7c8
  loc:	%l0      0x60105ec5380	%l1         0x2b438000	
	%l2               0x10	%l3          0xc2422e0	
	%l4                  0	%l5          0x19bd000	
	%l6             0x19bd	%l7      0x60086a0cf80	
  in:	%i0      0x2a106cba710	%i1                  0	
	%i2             0x1800	%i3                0xf	
	%i4 0xffffffffffffffff	%i5               0x10	
	%fp      0x2a106cb9e51	%i7         0x7bf3dc18	
vxio:vol_splitsio_iogen+0xd4(0x2a106cba710, 0x0, 0x0, 0x7bf3dd68, 0x0, 
0x0)
vxio:vol_subdisk_iogen_base+0x1a8()
vxio:vol_subdisk_iogen(0x2a106cba810, 0x2a106cbaa38) - frame recycled
vxio:volobject_iogen+0x34(0x2a106cba920, 0x2a106cbaa38)
vxio:volkiostart+0xa7c(0x60105ec5380, 0x2a106cbaa38, 0x0)

 

The Operating System encounters a panic() because of the incorrect use of device numbers. Volume Manager uses the whole slice number s0 instead of s7 which represents the whole
device for the EFI label of the disk. Due to this, the device open fails and the incorrect disk 'maxiosize' gets populated. While performing an I/O operation, the operating system panics with divide by zero error.

Veritas has fixed the code in SF HA 6.0.3 so that Volume Manager now correctly uses the device number corresponding to partition 7 of the device if the device size is greater than or equal to 1TB.

The fix for this issue can be downloaded from our SORT patch portal or directly via the below links:

sfha-sol10_sparc-6.0.3 - https://sort.veritas.com/patch/detail/7029
sfha-sol11_sparc-6.0.3 - https://sort.veritas.com/patch/detail/7030
sfha-sol10_x64-6.0.3 - https://sort.veritas.com/patch/detail/7031
sfha-sol11_x64-6.0.3 - https://sort.veritas.com/patch/detail/7032

 

 

Applies To

The issue is only encountered when the disk on which the resize operation is being performed has an EFI format or has SMI label where the "New" disk size after the resize operation is over 1TB.

Note: All devices larger than or equal to 1TB require to have an EFI label as per Solaris operating system requirement.

 

If an existing EFI labeled device or a SMI labeled device on a Solaris operating system that is being resized to a new size that will be larger than 1TB, then a "vxdisk resize" operation in Volume Manager to resize the device to the new size can lead to an operating system panic. This issue could occur if either of the the following two conditions are true: 1.) If resizing a CDS disk from less than 1TB to >= 1TB on a Solaris Operating System OR 2.) If resizing a CDS disk that is already >= 1TB and the new size is also >=1TB