Seeing new EMC disks as 0-bytes

[S.Rider]

We recently attached some EMC Clarion disks to an rp4440 running 11.23. Most (but not all) of the EMC disks are showing as 0-bytes but some show okay. Here's a sample diskinfo.

disk is /dev/rdsk/c13t0d2
SCSI describe of /dev/rdsk/c13t0d2:
vendor: DGC
product id: CX700WDR5
type: direct access
size: 0 Kbytes
bytes per sector: 0
-----------------------------
disk is /dev/rdsk/c13t0d3
SCSI describe of /dev/rdsk/c13t0d3:
vendor: DGC
product id: CX700WDR5
type: direct access
size: 37764096 Kbytes
bytes per sector: 512

If I try to vgcreate using one of the 0-byte luns, it dies "Device Busy". Also, anything like an ioscan or a diskinfo that goes against those 0-byte luns takes minutes. Any ideas ?

[Alex Lavrov.]

I have a strong feeling that something went wrong during the creation of the luns. Did you review all the settings in the EMC administration tools?

Alex.

[Raj D.]

Hi Jay ,

Try cross checking with sym commands.

#sympd list | grep dev/rdsk/c13t0d2
You will get the output size like this.

--------------------------------
# sympd list | more

Symmetrix ID: 000187720590

Device Name Directors Device
--------------------------- ------------- -------------------------------------
Cap
Physical Sym SA :P DA :IT Config Attribute Sts (MB)
--------------------------- ------------- -------------------------------------

/dev/rdsk/c73t0d0 0000 10A:0 15D:DE 2-Way Mir N/Grp'd VCM WD 23
/dev/rdsk/c73t3d7 043D 10A:0 15B:DA 2-Way Mir N/Grp'd (M) RW 34526
/dev/rdsk/c73t4d0 093D 10A:0 15A:C0 2-Way Mir N/Grp'd RW 8632
/dev/rdsk/c73t4d1 093E 10A:0 16D:C0 2-Way Mir N/Grp'd RW 8632
/dev/rdsk/c73t4d2 093F 10A:0 15D:C3 2-Way Mir N/Grp'd RW 8632
/dev/rdsk/c73t5d1 0465 10A:0 01A:C0 2-Way Mir N/Grp'd (M) RW 34526
/dev/rdsk/c73t5d2 0469 10A:0 15D:C3 2-Way Mir N/Grp'd (M) RW 34526
/dev/rdsk/c73t5d4 0471 10A:0 01C:C2 2-Way Mir N/Grp'd (M) RW 34526
/dev/rdsk/c73t5d5 0475 10A:0 01A:C2 2-Way Mir N/Grp'd (M) RW 34526
/dev/rdsk/c73t5d6 0479 10A:0 15D:C5 2-Way Mir N/Grp'd (M) RW 34526
/dev/rdsk/c73t5d7 047D 10A:0 15B:C5 2-Way Mir N/Grp'd (M) RW 34526
/dev/rdsk/c73t6d0 0481 10A:0 01C:C4 2-Way Mir N/Grp'd (M) RW 34526
/dev/rdsk/c73t6d1 0485 10A:0 01A:C4 2-Way Mir N/Grp'd (M) RW 34526
/dev/rdsk/c73t6d3 004E 10A:0 16A:C7 2-Way Mir N/Grp'd RW 4647
/dev/rdsk/c73t6d4 0051 10A:0 15B:C7 2-Way Mir N/Grp'd (M) RW 18589
/dev/rdsk/c73t6d5 0055 10A:0 01C:C6 2-Way Mir N/Grp'd (M) RW 9294
-------------------------------------------

Last column all says the size, i.e 34526 is 36GB disk , and 8632 is 9GB disk ,

if still shows zero , it seems problem with creating the Lun or may be the disk is having problem.

Hope this may help you,

Raj.

[Steven E. Protter]

I have seen the disk arrays themselves be misdetected as disks when no disk is assigned to LUN0.

They look just like what you report. Or it could be the post before mine is right.

Take a look with mstm or xstm or cstm and you might get an idea which it is.

I'd like to seee the ioscan output if it weren't past my bedtime. Still, it would be useful to verify my quess.

SEP

[Tim Nelson]

symdev show devid

i.e.
symdev show 01F

will give full info on the volume as well

[Florian Heigl (new acc)]

symdef will IMHO not show very much for CX luns, or rather not be installed at all. :)

I'd hope a navicli getlun is the command in question.

[Ashwani Kashyap]

IF you have "inq" installed on the host , then you can run
#inq -clariion
and
#inq -clar_wwn
to check if you actually correctly see the devices and they are logged in properly with the SP's .

If not then deregister the host from the clariion and remove the LUN's from the storage group and destroy the storage group .Then reregister the host and put back the storage group and the LUNs in it . Some times it helps . YOur apps have to come down though .