SAN disks on Solaris in error state and cannot get initialised

[Maveric63]

Hi

I zoned some SAN disks to a Solaris 10 host but am unable to initialise the devices. The status is error 

#VXDISK list output

bash-3.00# vxdisk list
DEVICE       TYPE            DISK         GROUP        STATUS
disk_0       auto:cdsdisk    data01       REGAN_DG     online
disk_1       auto:none       -            -            online invalid
disk_2       auto:cdsdisk    data02       REGAN_DG     online
disk_3       auto:simple     -            -            online invalid
emc1_3a2a    auto            -            -            error
emc1_3a2b    auto            -            -            error
emc1_3a2c    auto            -            -            error
emc1_3a2d    auto            -            -            error
emc1_3a2e    auto            -            -            error
emc1_3a2f    auto            -            -            error
emc1_3a3a    auto            -            -            error

When trying to do vxdisksetup

#vxdisksetup -i emc1_3a2a

VxVM vxdisksetup ERROR V-5-2-3535 emc1_3a2a: Invalid dmpnodename for disk device emc1_3a2a.

Any ideas how to get these disks online

 

[Gaurav_S]

what is the version of  OS & vxvm ?

did u label the devices from OS using format command ?

Seems like devices are inaccessible,  can u paste following outputs:

# vxdisk -e list

# modinfo | grep -i vx

# vxdisk list emc1_3a2a

Gaurav

[Maveric63]

Hi Gaurav

OS Version

Solaris 5.10

Veritas version

 PKGINST:  VRTSvxvm
      NAME:  Binaries for VERITAS Volume Manager by Symantec
  CATEGORY:  system
      ARCH:  sparc
   VERSION:  5.1,REV=10.06.2009.22.05
   BASEDIR:  /
    VENDOR:  Symantec Corporation
      DESC:  Virtual Disk Subsystem
    PSTAMP:  5.1.100.000-5.1SP1-2010-10-10
  INSTDATE:  Feb 23 2012 13:00
   HOTLINE:  http://support.veritas.com/phonesup/phonesup_ddProduct_.htm
     EMAIL:  support@veritas.com
    STATUS:  completely installed
     FILES:      951 installed pathnames
                  41 shared pathnames
                 115 directories
                 425 executables
              417704 blocks used (approx)
 

Doing format for OS does not work

 format
Ready to format.  Formatting cannot be interrupted.
Continue? yes
Beginning format. The current time is Wed Jun  6 11:10:48 2012

Reserve failed
failed
format> ^C

Warning: Current geometry overshoots actual geometry of disk

Continue labelling disk? yes
Warning: error writing VTOC.
Not ready error during read
ASC: 0x4   ASCQ: 0x3
Warning: error reading backup label.
Not ready error during read
ASC: 0x4   ASCQ: 0x3
Warning: error reading backup label.
Not ready error during read
ASC: 0x4   ASCQ: 0x3
Warning: error reading backup label.
Not ready error during read
ASC: 0x4   ASCQ: 0x3
Warning: error reading backup label.
Not ready error during read
ASC: 0x4   ASCQ: 0x3
Warning: error reading backup label.
Warning: no backup labels

#vxdisk -e list output

emc1_3a2a    auto           -            -           error                c0t500009740800B55Dd7s2 bcv-nr
emc1_3a2b    auto           -            -           error                c0t500009740800B55Dd8s2 bcv-nr
emc1_3a2c    auto           -            -           error                c0t500009740800B55Dd9s2 bcv-nr
emc1_3a2d    auto           -            -           error                c0t500009740800B55Dd10s2 bcv-nr
emc1_3a2e    auto           -            -           error                c0t500009740800B55Dd11s2 bcv-nr
emc1_3a2f    auto           -            -           error                c0t500009740800B55Dd12s2 bcv-nr
emc1_3a3a    auto           -            -           error                c0t500009740800B55Dd23s2 bcv-nr
emc1_3a3b    auto           -            -           error                c0t500009740800B55Dd24s2 bcv-nr
emc1_3a3c    auto           -            -           error                c0t500009740800B55Dd25s2 bcv-nr
emc1_3a4c    auto           -            -           error                c0t500009740800B55Dd29s2 bcv-nr
emc1_3a5c    auto           -            -           error                c0t500009740800B55Dd33s2 bcv-nr
emc1_3a6c    auto           -            -           error                c0t500009740800B55Dd37s2 bcv-nr
emc1_3a7d    auto           -            -           error                c0t500009740800B55Dd42s2 bcv-nr
emc1_3a29    auto           -            -           error                c0t500009740800B55Dd6s2 bcv-nr
emc1_3a30    auto           -            -           error                c0t500009740800B55Dd13s2 bcv-nr
emc1_3a31    auto           -            -           error                c0t500009740800B55Dd14s2 bcv-nr
emc1_3a32    auto           -            -           error                c0t500009740800B55Dd15s2 bcv-nr
emc1_3a33    auto           -            -           error                c0t500009740800B55Dd16s2 bcv-nr
 

#modinfo output

 modinfo | grep -i vx
 30  1372120  50af8 334   1  vxdmp (VxVM 5.1SP1 DMP Driver)
 32 7be00000 21a630 335   1  vxio (VxVM 5.1SP1 I/O driver)
 34  13bc290   1190 336   1  vxspec (VxVM 5.1SP1 control/status driv)
239 7ba0d208    d40 337   1  vxportal (VxFS 5.1_SP1 portal driver)
240 7a600000 1fb718  21   1  vxfs (VxFS 5.1_SP1 SunOS 5.10)
253 7ab24000   b7a0 338   1  fdd (VxQIO 5.1_SP1 Quick I/O driver)
 

#vxdisk list ouput

vxdisk list emc1_3a2a
Device:    emc1_3a2a
devicetag: emc1_3a2a
type:      auto
flags:     error private autoconfig
pubpaths:  block=/dev/vx/dmp/emc1_3a2a char=/dev/vx/rdmp/emc1_3a2a
guid:      -
udid:      EMC%5FSYMMETRIX%5F000292600045%5F4503A2A000
site:      -
errno:     Configuration daemon error -1
Multipathing information:
numpaths:   1
c0t500009740800B55Dd7s2 state=enabled
 

[Gaurav_S]

Luns looks to be BCV split luns ... however looking at above I believe there is no access to the Luns, that is the reason you are getting a write error even when trying a format command ...

From this, I can atleast conclude that problem is not at veritas layer rather it is at one layer below veritas where you need to check if luns are presented correctly with read/write access ...

are you able to read the VTOC ?

# prtvtoc /dev/rdsk/c0t500009740800B55Dd42s2

how does the split of luns happened ? & how are these luns presented here ?

Does the source machine of source luns use IOfencing or any sort of reservation mechanism at SAN layer ?

G

[Maveric63]

Ok I found a problem and way to get the disks out of the error state I think.
On the SAN all the devices were set to DEVICE NOT READY

I changed this parameter on all to READY.

Nowwhen I run the vxdisk -e list it does not respond with any output, just returns me to the comand prompt.
However  now when I run the Solaris FORMAT command the illegal device entires are missing

Checking why SF is not detecting devices now

The BCV /Split is managed by an EMc application called Replication Manager

[Gaurav_S · Accepted solution]

Check out device tree cleanup technote, there could be a possibility of device tree having issues .. you can perform a cleanup by following link:

http://www.symantec.com/docs/TECH78463

If not working, probably restart the server with reconfig option would be my last go....

G

[Marianne]

Please also try 'vxdisk scandisks'

Have you been able to label the disks/luns?

What are your lun sizes?

[Maveric63]

Hi

Restarted server and all seems Ok.

I can see all my diskgroups and EMC Replication Manager can now mount all replicated copies

Thanks for the help all