pvcreate: IO error on Physical Volume device "/dev/rdsk/c4t0d1".

[Lee Hallin]

I have an EVA3000 that I recently configured a new host on (it had 3 previously).  I "present" a VDISK to the new host.  The present "seems" to work fine.  No errors and it says it is "presented to <newhost>.

When I go to the host and do an "ioscan -fnC disk", I see the newLun and device files listed.

When I do a "pvcreate /dev/rdsk/c4t0d1", I get 'pvcreate: IO error on Physical Volume device "/dev/rdsk/c4t0d1".'

If I include the -f (force) flag, I get:

pvcreate: Error: Couldn't read LVMREC for the physical volume "/dev/rdsk/c4t0d1".

The new host is running hp-ux 11.31 whereas the other three are running 11i.  Could the 11.31 be an issue?

Thanks for any light you might be able to shed :-)

Lee

[chris huys_4]

Its 11.31, so you should use the /dev/rdisk/diskX devicefile, with pvcreate, instead of a legacy device file.

If that doesnt work.

# ioscan -fnN

# scsimgr get_info -D /dev/rdisk/diskX

# ioscan -m dsf

[Lee Hallin]

Thanks Chris,

Looks like the new device is "disk8" (according to the following which, to me, equates disk8 to c4t0d1) so I replaced the /dev/rdsk/c4t0d1 with /dev/rdisk/disk8 and got basically the same error.

Do I need to re-boot the host at any particular point?

LASR24 root / # ioscan -m dsf
Persistent DSF           Legacy DSF(s)
========================================
/dev/rdisk/disk2         /dev/rdsk/c0t0d0
/dev/rdisk/disk2_p1      /dev/rdsk/c0t0d0s1
/dev/rdisk/disk2_p2      /dev/rdsk/c0t0d0s2
/dev/rdisk/disk2_p3      /dev/rdsk/c0t0d0s3
/dev/rdisk/disk3         /dev/rdsk/c1t0d0
/dev/pt/pt5              /dev/rscsi/c4t0d0
                         /dev/rscsi/c6t0d0
                         /dev/rscsi/c8t0d0
                         /dev/rscsi/c10t0d0
/dev/rdisk/disk8         /dev/rdsk/c4t0d1
                         /dev/rdsk/c6t0d1
LASR24 root / # date;pvcreate /dev/rdisk/disk8;date  
Tue Aug  2 17:03:01 PDT 2011
pvcreate: IO error on Physical Volume device "/dev/rdisk/disk8".
Tue Aug  2 17:03:11 PDT 2011
LASR24 root / #

Thanks for your help,

Lee

[chris huys_4]

scsimgr get_info -D /dev/rdisk/disk8

scsimgr lun_map -D /de/vrdisk/disk8

[Lee Hallin]

Thanks Chris,

  Below are executions of the two commands you asked for (i think):

LASR24 root / # date
Tue Aug  2 17:54:28 PDT 2011
LASR24 root / # scsimgr get_info -D /dev/rdisk/disk8

        STATUS INFORMATION FOR LUN : /dev/rdisk/disk8

Generic Status Information

SCSI services internal state                  = UNOPEN
Device type                                   = Direct_Access
EVPD page 0x83 description code               = 1
EVPD page 0x83 description association        = 0
EVPD page 0x83 description type               = 3
World Wide Identifier (WWID)                  = 0x600508b40001390a00018000000b0000
Serial number                                 = "P66C5D4AAPX016"
Vendor id                                     = "HP      "
Product id                                    = "HSV100          "
Product revision                              = "2001"
Other properties                              = ""
SPC protocol revision                         = 4
Open count (includes chr/blk/pass-thru/class) = 0
Raw open count (includes class/pass-thru)     = 0
Pass-thru opens                               = 0
LUN path count                                = 2
Active LUN paths                              = 2
Standby LUN paths                             = 0
Failed LUN paths                              = 0
Maximum I/O size allowed                      = 2097152
Preferred I/O size                            = 2097152
Outstanding I/Os                              = 0
I/O load balance policy                       = round_robin
Path fail threshold time period               = 0
Transient time period                         = 120
Tracing buffer size                           = 1024
LUN Path used when policy is path_lockdown    = NA
LUN access type                               = NA
Asymmetric logical unit access supported      = No
Asymmetric states supported                   = NA
Preferred paths reported by device            = No
Preferred LUN paths                           = 0

Driver esdisk Status Information :

Capacity in number of blocks                      = 146800640
Block size in bytes                               = 512
Number of active IOs                              = 0
Special properties                                =
Maximum number of IO retries                      = 45
IO transfer timeout in secs                       = 30
FORMAT command timeout in secs                    = 86400
START UNIT command timeout in secs                = 60
Timeout in secs before starting failing IO        = 120
IO infinite retries                               = false
LASR24 root / # scsimgr lun_map -D /dev/rdisk/disk8

        LUN PATH INFORMATION FOR LUN : /dev/rdisk/disk8

Total number of LUN paths     = 2
World Wide Identifier(WWID)    = 0x600508b40001390a00018000000b0000

LUN path : lunpath7
Class                         = lunpath
Instance                      = 7
Hardware path                 = 0/0/0/9/0/0/0.0x50001fe150032a29.0x1000000000000
SCSI transport protocol       = fibre_channel
State                         = UNOPEN
Last Open or Close state      = ACTIVE

LUN path : lunpath6
Class                         = lunpath
Instance                      = 6
Hardware path                 = 0/0/0/9/0/0/0.0x50001fe150032a2c.0x1000000000000
SCSI transport protocol       = fibre_channel
State                         = UNOPEN
Last Open or Close state      = ACTIVE

LASR24 root / #

[chris huys_4 · Accepted solution]

You have a very old, not updated, EVA3000. If im not mistaken, HP-UX 11.31, requires that you update the EVA3000 to the vcs 4.x firmware, so that the EVA supports the active/active asymmetric "protocol".. I think your EVA3000 has still the older vcs 3.x firmware..

Update!! ;)

Greetz,
Chris 

[Lee Hallin]

Thanks Chris,

 Yes, updating the firmware was on the to-do list already!  I just didn't realize that 11.31 REQUIRED the upgrade :-)

Thank you so very much for your help.  You have been very helpful :-)

Sincerely,  Lee

[Torsten.]

There are 2 important points:

- array must be active/active (depends on firmware version)

- host mode must be set to "hp-ux"

[Lee Hallin]

Thanks Torsten,

  I read your #2 item and said hmmm, I don't remember specifying HP-UX so I went and checked and, oops, it was set to Windows (what a PANE :-) so I un-presented it, change it from Windows to HP-UX.  Then I re-presented it and tried the pvcreate on HP-UX and still got same error.  Then, for grins, I un-presented it and re-presented it to a hp-ux 11i (instead of 11.31) box and the pvcreate worked fine!  This means the presentation is working; only the OS is an issue with current version of switch firmware.  Plan is to fix that tomorrow morning.

Thanks for your post