11.31 install woes

[Bill Hassell]

After cold installing 11.31 latest version and discovering the networking was locked down by IPfilter, I am now facing disabled disks, I have 4 JBOD SCSI disks, two of which were included in creating vg00. I now find that the other two (which were working just fine) now report zero size with diskinfo. After poking around in SMH, I found the new cprop command in /opt/proplus which reports this for the disabled disks:

[Instance]: 3
****************************************************
[Hash ID]: Disk Drive:900416491
[Status]: Degraded
[Legacy Hardware Path]: 0/3/1/0.4.0
[Agile Hardware Path]: 64000/0xfa00/0x8
[Product Id]: ST373454LC
[Vendor Id]: HP 73.4G
[Serial Number]: 3KP1H98L00007611V91L
[Capacity (GB)]: 0.0
[Disk FW Revision]: HPC3
****************************************************

# diskinfo /dev/rdsk/c6t4d0
SCSI describe of /dev/rdsk/c6t4d0:
vendor: HP 73.4G
product id: ST373454LC
type: direct access
size: 0 Kbytes
bytes per sector: 0
rev level: HPC3
blocks per disk: 0
ISO version: 0
ECMA version: 0
ANSI version: 3
removable media: no
response format: 2
...


Needless to say, pvcreate and dd fail because the disk is 0 bytes in size. The disks are certainly not degraded (no errors in syslog) and were working 2 days ago with 11.23. I guessed that scsimgr enable might help:

# scsimgr enable -D /dev/rdisk/disk10
scsimgr: LUN /dev/rdisk/disk10 enabled successfully

but still size=0 with diskinfo. The disks are claimed in ioscan and the vendor ID shows at the end of the line.

Anyone else had these kinds of problems with the latest 11.31?

[James R. Ferguson · Accepted solution]

Hi Bill:

Is the firmware current enough for 11.31?

Perchance did you run the 'msv2v3check' tool prior to your install/upgrade?

Regards!

...JRF...

[Steven E. Protter]

Shalom Bill,

firmware is a definite possible problem. If you identify the server with the model command, it might be helpful in finding a solution.

I'm going to assume due to our off itrc communication that this was not an upgrade-ux install, it was indeed a cold install.

We are a little conservative and have not done any cold installs with the latest, greatest 11.31 release.

SEP

[Bill Hassell]

Looks like firmware but it is the SCSI Ultra320 plus Online Diags. The one condition I did not know about was that a mediainit was done on the two disks. If the Online Diags are running, it triggers a bug in the firmware that aborts the mediainit and destroys the disk. Normal measures cannot repair this damage -- cstm expert mode must be used to reformat the disks.

Here are the details, page 4:
http://docs.hp.com/en/5900-0113/5900-0113.pdf

[Bill Hassell]

This document describes the condition:

http://docs.hp.com/en/5900-0113/5900-0113.pdf