Super Dome 2 and Ignite Cloning

[lzsilva]

Super Dome 2 and Ignite Cloning

Good afternoon,

We have here a HP-UX 11.31 (B.11.31.0803 ) nPar on a Super Dome SX2000 with local disks for the OS filesystems.

I’ve made a network ignite of this server to clone it to a Superdome2 16s (where the OS filesystems are in a external P2000).

The ignite restore ends with just one warning an no errors:

insf: Installing special files for pseudo driver kepd

insf: Installing special files for pseudo driver framebuf

insf: Installing special files for pseudo driver sad

       * Creating the system file.

Creating your system file in the file /tmp/ign_configure/tstand.system

       * Constructing the bootconf file.

       * Setting primary boot path to "/dev/rdisk/disk1".

Primary boot path set to 33/0/2/0/0/0/0.0x207000c0ff11af65.0x4000000000000000 (/dev/rdisk/disk1)

       * Executing: "/var/adm/sw/products/OS-Core/CORE-SHLIBS/iux_postload".

       * Running: /var/adm/sw/products/OS-Core/CORE-SHLIBS/postinstall

       * Executing: "/var/adm/sw/products/OS-Core/CORE2-KRN/iux_postload".

       * Executing: "/var/adm/sw/products/OS-Core/CORE-KRN/iux_postload".

       * Executing: "/var/adm/sw/products/FC-FCD/FC-FCD-KRN/iux_postload".

       * Executing: "/var/adm/sw/products/PHKL_40317/pfiles/iux_postload".

       * Running: /var/adm/sw/products/PHKL_40317/pfiles/postinstall

       * Executing: "/var/adm/sw/products/PHKL_41553/pfiles/iux_postload".

       * Backing up LVM configuration for "vg00".

       * Loading_software:  Complete

=======  10/20/11 14:11:19 BDT  Building the new kernel.

       * Build_Kernel:  Begin

       * Building a new kernel using the command: "/usr/sbin/mk_kernel -f -o /stand/vmunix".

WARNING: The configuration is supposed to include a module 'iexgbe' that

         is not installed on the system.  The system will attempt to

         proceed without it, but the resulting kernel may fail to boot

         or function properly.

       * The currently running configuration has been imported from

         '/stand/system'.

       * Build_Kernel:  Complete

       * Boot_From_Client_Disk:  Begin

=======  10/20/11 14:11:55 BDT  Rebooting machine as expected.

NOTE:    Rebooting system.

Calling function e0000000020ec140 for Shutdown State 1 type 0x1

But the new server boots up with kernel panic:

Console is on Serial Device - via PCDP

Booting kernel...

Stored message buffer up to system crash:

/0 pci_slot

5/0/1/2/0 PCItoPCI

5/0/1/2/0/0 pci_slot

5/0/2 gh2p

5/0/2/0/0 PCItoPCI

5/0/2/0/0/0 pci_slot

5/0/2/2/0 PCItoPCI

5/0/2/2/0/0 pci_slot

5/0/2/2/0/0/0 PCItoPCI

5/0/2/2/0/0/0/1/0 hcd

5/0/2/2/0/0/0/1/1 hcd

5/0/2/2/0/0/0/1/2 ehci

5/0/2/3/0 PCItoPCI

5/0/2/3/0/0 pci_slot

5/0/2/3/0/0/0 sysdev

5/0/2/3/0/0/2 sysdev

5/0/2/3/0/0/4 uhci

5/120 processor

5/122 processor

5/124 processor

5/126 processor

5/128 processor

5/130 processor

5/132 processor

5/134 processor

6 cell

6/0 cec_gen

6/0/0 gh2p

6/0/0/0/0 PCItoPCI

6/0/0/0/0/0 pci_slot

6/0/0/2/0 PCItoPCI

6/0/0/2/0/0 pci_slot

6/0/1 gh2p

6/0/1/0/0 PCItoPCI

6/0/1/0/0/0 pci_slot

6/0/1/2/0 PCItoPCI

6/0/1/2/0/0 pci_slot

6/0/2 gh2p

6/0/2/0/0 PCItoPCI

6/0/2/0/0/0 pci_slot

6/0/2/2/0 PCItoPCI

6/0/2/2/0/0 pci_slot

6/0/2/2/0/0/0 PCItoPCI

6/0/2/2/0/0/0/1/0 hcd

6/0/2/2/0/0/0/1/1 hcd

6/0/2/2/0/0/0/1/2 ehci

6/0/2/3/0 PCItoPCI

6/0/2/3/0/0 pci_slot

6/0/2/3/0/0/0 sysdev

6/0/2/3/0/0/2 sysdev

6/0/2/3/0/0/4 uhci

6/120 processor

6/122 processor

6/124 processor

6/126 processor

6/128 processor

6/130 processor

6/132 processor

6/134 processor

7 cell

7/0 cec_gen

7/0/0 gh2p

7/0/0/0/0 PCItoPCI

7/0/0/0/0/0 pci_slot

7/0/0/2/0 PCItoPCI

7/0/0/2/0/0 pci_slot

7/0/1 gh2p

7/0/1/0/0 PCItoPCI

7/0/1/0/0/0 pci_slot

7/0/1/2/0 PCItoPCI

7/0/1/2/0/0 pci_slot

7/0/2 gh2p

7/0/2/0/0 PCItoPCI

7/0/2/0/0/0 pci_slot

7/0/2/2/0 PCItoPCI

7/0/2/2/0/0 pci_slot

7/0/2/2/0/0/0 PCItoPCI

7/0/2/2/0/0/0/1/0 hcd

7/0/2/2/0/0/0/1/1 hcd

7/0/2/2/0/0/0/1/2 ehci

7/0/2/3/0 PCItoPCI

7/0/2/3/0/0 pci_slot

7/0/2/3/0/0/0 sysdev

7/0/2/3/0/0/2 sysdev

7/0/2/3/0/0/4 uhci

7/120 processor

7/122 processor

7/124 processor

7/126 processor

7/128 processor

7/130 processor

7/132 processor

7/134 processor

33 cell

33/0 cec_gen

33/0/0 gh2p

33/0/0/0/0 PCItoPCI

33/0/0/0/0/0 pci_slot

33/0/0/2/0 PCItoPCI

33/0/0/2/0/0 pci_slot

33/0/1 gh2p

33/0/1/0/0 PCItoPCI

33/0/1/0/0/0 pci_slot

33/0/1/2/0 PCItoPCI

33/0/1/2/0/0 pci_slot

33/0/2 gh2p

33/0/2/0/0 PCItoPCI

33/0/2/0/0/0 pci_slot

33/0/2/2/0 PCItoPCI

33/0/2/2/0/0 pci_slot

255/1 mass_storage

255/1/0 usb_ms_scsi

gio: Error configuring root device using EFI hw paths

    System Console is on the Built-In Serial Interface

AF_INET socket/streams output daemon running, pid 135

afinet_prelink: module installed

Starting the STREAMS daemons-phase 1

    Swap device table:  (start & size given in 512-byte blocks)

        entry 0 - auto-configured on root device; 6/0/2/3/0/0/4.1 usbcomposite

6/0/2/3/0/0/4.1.0 hid

6/0/2/3/0/0/4.1.1 hid

5/0/2/3/0/0/4.1 usbcomposite

5/0/2/3/0/0/4.1.0 hid

5/0/2/3/0/0/4.1.1 hid

4/0/2/3/0/0/4.1 usbcomposite

4/0/2/3/0/0/4.1.0 hid

4/0/2/3/0/0/4.1.1 hid

2/0/2/3/0/0/4.1 usbcomposite

2/0/2/3/0/0/4.1.0 hid

2/0/2/3/0/0/4.1.1 hid

0/0/2/3/0/0/4.1 usbcomposite

0/0/2/3/0/0/4.1.0 hid

0/0/2/3/0/0/4.1.1 hid

7/0/2/3/0/0/4.1 usbcomposite

7/0/2/3/0/0/4.1.0 hid

7/0/2/3/0/0/4.1.1 hid

ignored - can't open device

WARNING: No swap device configured, so dump cannot be defaulted to primary swap.

WARNING: No dump devices are configured.  Dump is disabled.

mountfs:opend

VFS_MOUNTROOT error information:

0:(hfs) 6

1:(nfs) 223

9:(vxfs) 6

10:(nfs3) 223

11:(nfs4) 223

12:(autofs) 251

14:(cifs) 223

17:(dnlc_neg_fs) 251

Message buffer contents after system crash:

panic: VFS_MOUNTROOT: READ_ONLY: all VFS_MOUNTROOTs failed: NEED DRIVERS ?????

Stack Trace:

  IP                  Function Name

  0xe000000001f2d0c0  vfs_mountroot+0x700

  0xe00000000160ff80  im_mountroot+0x40

  0xe000000001f1de20  DoCalllist+0x390

End of Stack Trace

It is quite obvious that there is some kernel module missing in the installation, but how can I find witch module is missing?

Regards,

Lincoln

[James R. Ferguson]

Hi:

I'd try installing PHKL_41368 and then taking another Ignite image for the clone.  See the patch notes for the aforementioned.

Regards!

...JRF...

[Stephan.]

I would go with a golden Image instead of the recovery to clone it, using the golden image will allow the installation process to install missing drivers and/or kernel modules.

hth,

Stephan

[lzsilva]

Thanks James,

I'll try it...

Stephan,

Thanks for your reply, but what do you mean with "golden Image"?

[ClaroBR]

James,

This patch is already installed:

server@root:/root # swlist -l product |grep -i "SCSI cumulative"
  PHKL_41368            1.0            SCSI cumulative I/O patch

[Stephan.]

I guess you used make_net_recovery and the procedure of the whitepaper "successful system cloning using Ignite/UX".

This is fine as long as you stay on the same platform and hardware, but you are changing a lot of hardware.

So use make_sys_image to create a golden image, configure it on the ignite server and install from it, this will not be a 100% clone. It is a complete new install but will keep most of the customizations.

Check http://docstore.mik.ua/manuals/hp-ux/en/5992-5309/ch11.html for details.

hth,

Stephan

[lzsilva]

Stephan,

Unfortunately I will not be able to use golden image because of the following note in the documentation:

IMPORTANT: Do not use the system while make_sys_image is running in the next step. Device files are removed, and the host and/or networking information on the system is reset. After the command is complete, these files are put back.

The source server is a production server…

[Stephan.]

:-/

The usual problem ... as mentioned it would be a lot more safe and perhaps you can arrange a downtime. Otherwise I would go with a cold install of the current release, everything else will cause in my opinion more trouble then producing any advantage.

Best,

Stephan

[Torsten.]

> We have here a HP-UX 11.31 (B.11.31.0803 )

This is much too old.

For the current SD2 firmware you must have version 1109 (september 2011).

But even without the newest firmware the minimum OS release for Superdome 2 is one of the 2010 releases, AFAIR - 2008 is definetely much too old to support this hardware!

[Duncan Edmonstone]

Torsten is correct - the minimum HP-UX11iv3 update release to run Superdome 2 is Sept 2010 (1009), and that is good for everthing up to and including SD2 firmware release 1.3.1. If you are on the absolute latest firmware (2.2.27) you will need to install the latest release of HP-UX11iv3 - Sept 2011 (1109) and also patch PHKL_42230

List of OS update and patches required (for the older firmware) is here:

http://bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c02468581/c02468581.pdf