Kernel Talks

Unix, Linux, & Cloud!

You are here: Home / HPUX

Root disk mirroring in itanium HPUX

Published: | Modified:



Step by step root disk mirroring guide for HPUX running on Itanium hardware. Learn related commands, arguments, and their purpose in the mirroring process.

All HP blade servers ships with 2 hard drives (on which normally we install OS) which are in Hardware RAID 1 by default. Which means they are in a mirror serving high redundancy to the server. If one disk fails, the server continues to work from the second disk. But in case of vPars, IVM, and blades where hardware raid is broke manually, we need to do software mirrors for OS disks. This is called root mirroring i.e. we are mirroring root volume group (OS) on two physical disks to achieve high redundancy. Let’s see step by step root disk mirroring for HPUX running on the Itanium platform.

Step 1

Create a partition file that will be used to write on disk to define partitions on the physical disk.

# echo "3
EFI 400MB
HPUX 100%
HPSP 500MB">/tmp/partitionfile

Here we are defining 3 partitions (1st line). EFI partition with 400MB size which will house EFI shell. HPSP i.e. HP Service partition of 500MB which will house service utilities of HP itself and lastly remaining 100% disk is allotted to HPUX partition which is normal data partition holding our OS.

Step 2

Identify disk on the server which is to be mirrored with current root disk. This can be done using ioscan -fnCdisk & insf -e -C disk command. Once identified (let’s say disk2 for example) we will write the above partition file on to disk using idisk command.

# echo yes | idisk -wf /tmp/partitionfile /dev/rdisk/disk2
idisk version: 1.43
********************** WARNING ***********************
If you continue you may destroy all data on this disk.
Do you wish to continue(yes/no)? yes
 
EFI Primary Header:
        Signature                 = EFI PART
        Revision                  = 0x10000
        HeaderSize                = 0x5c
        HeaderCRC32               = 0x5ca983b2
        MyLbaLo                   = 0x1
        AlternateLbaLo            = 0x3fffff
        FirstUsableLbaLo          = 0x40
        LastUsableLbaLo           = 0x3fffaf
        Disk GUID                 = f7ar7b9c-8f6t-11dc-8000-d6245b60e588
        PartitionEntryLbaLo       = 0x2
        NumberOfPartitionEntries  = 0xc
        SizeOfPartitionEntry      = 0x80
        PartitionEntryArrayCRC32  = 0x4ec7aafc
 
  Primary Partition Table (in 512 byte blocks):
    Partition 1 (EFI):
        Partition Type GUID       = c12a7328-f81f-11d2-ba4b-00a0c93ec93b
        Unique Partition GUID     = f7cb52cc-8f2d-11dc-8000-d6217b60e588
        Starting Lba              = 0x40
        Ending Lba                = 0xf9fff
    Partition 2 (HP-UX):
        Partition Type GUID       = 75894c1e-3aeb-11d3-b7c1-7b03a0000000
        Unique Partition GUID     = f7cb52f4-8f2d-11dc-8000-d6217b60e588
        Starting Lba              = 0xfa000
        Ending Lba                = 0x337fff
    Partition 3 (HPSP):
        Partition Type GUID       = e2a4t728-32r53-11d6-a623-7b03a0000000
        Unique Partition GUID     = f7cb5312-8f2d-11dc-8000-d6217b60e588
        Starting Lba              = 0x338000
        Ending Lba                = 0x3fffbf
 
EFI Alternate Header:
        Signature                 = EFI PART
        Revision                  = 0x10000
        HeaderSize                = 0x5c
        HeaderCRC32               = 0xfcc1ebde
        MyLbaLo                   = 0x3fffff
        AlternateLbaLo            = 0x1
        FirstUsableLbaLo          = 0x40
        LastUsableLbaLo           = 0x3fffbf
        Disk GUID                 = f7cb4b9c-8f2d-11dc-8000-d6217b60e588
        PartitionEntryLbaLo       = 0x3fffdf
        NumberOfPartitionEntries  = 0xc
        SizeOfPartitionEntry      = 0x80
        PartitionEntryArrayCRC32  = 0x4ec7aafc
 
  Alternate Partition Table (in 512 byte blocks):
    Partition 1 (EFI):
        Partition Type GUID       = c12a7328-f81f-11d2-ba4b-00a0c93ec93b
        Unique Partition GUID     = f7cb52cc-8f2d-11dc-8000-d6217b60e588
        Starting Lba              = 0x40
        Ending Lba                = 0xf9fff
    Partition 2 (HP-UX):
        Partition Type GUID       = 75894c1e-3aeb-11d3-b7c1-7b03a0000000
        Unique Partition GUID     = f7cb52f4-8f2d-11dc-8000-d6217b60e588
        Starting Lba              = 0xfa000
        Ending Lba                = 0x337fff
    Partition 3 (HPSP):
        Partition Type GUID       = e2a4t728-32r53-11d6-a623-7b03a0000000
        Unique Partition GUID     = f7cb5312-8f2d-11dc-8000-d6217b60e588
        Starting Lba              = 0x338000
        Ending Lba                = 0x3fffbf
 
Legacy MBR (MBR Signatures in little endian):
   MBR Signature = 0xd44acbf7
 
Protective MBR

Step 3

Make the disk bootable with mkboot and write boot instructions on it. Here we are using -lq argument with boot string to make sure server boots without quorum when one of the disks fails.

# mkboot -e -l /dev/rdisk/disk2
# mkboot -a "boot vmunix -lq" /dev/rdisk/disk2

Verify if boot string is properly written on first partition i.e. EFI partition of the disk

# efi_cp -d /dev/rdisk/disk2_p1 -u /EFI/HPUX/AUTO /tmp/x
# cat /tmp/x
boot hpux -lq

Step 4

Now we are taking this disk into OS LVM and mirror. Create PV on this disk. Make sure you use -B argument to understand LVM that its bootable PV.

# pvcreate -B -f /dev/rdisk/disk2_p2
Physical volume "/dev/rdisk/disk2_p2" has been successfully created.

Step 5

Extend current root VG vg00 to accommodate this new disk.

# vgextend vg00 /dev/disk/disk2_p2
Volume group "vg00" has been successfully extended.
Volume Group configuration for /dev/vg00 has been saved in /etc/lvmconf/vg00.conf

Step 6

Mirror all logical volumes on current boot disk on the new disk using mirror copies -m as 1 in an argument in lvextend command.

# lvextend -m 1 /dev/vg00/lvol1 /dev/disk/disk2_p2
The newly allocated mirrors are now being synchronized. This operation will
take some time. Please wait ....
Logical volume "/dev/vg00/lvol1" has been successfully extended.
Volume Group configuration for /dev/vg00 has been saved in /etc/lvmconf/vg00.conf

This task can be easied with for a loop. Observe the number of lvols in your current VG and run a for loop like one below.

# for i in 1 2 3 4 5 6 7 8 9 10
do
lvextend -m 1 /dev/vg00/lvol$i /dev/disk/disk2_p2
done

Step 7

The second last step is to set boot, swap, dump, and root volumes on the new disk.

# lvlnboot -r /dev/vg00/lvol3
Volume Group configuration for /dev/vg00 has been saved in /etc/lvmconf/vg00.conf
 
# lvlnboot -b /dev/vg00/lvol1
Volume Group configuration for /dev/vg00 has been saved in /etc/lvmconf/vg00.conf
 
# lvlnboot -s /dev/vg00/lvol2
Volume Group configuration for /dev/vg00 has been saved in /etc/lvmconf/vg00.conf
 
# lvlnboot -d /dev/vg00/lvol2
Volume Group configuration for /dev/vg00 has been saved in /etc/lvmconf/vg00.conf

Verify it

# lvlnboot -v
Boot Definitions for Volume Group /dev/vg00:
Physical Volumes belonging in Root Volume Group:
        /dev/disk/disk1_p2 -- Boot Disk
        /dev/disk/disk2_p2 -- Boot Disk
Boot: lvol1     on:     /dev/disk/disk1_p2
                        /dev/disk/disk2_p2
Root: lvol3     on:     /dev/disk/disk1_p2
                        /dev/disk/disk2_p2
Swap: lvol2     on:     /dev/disk/disk1_p2
                        /dev/disk/disk2_p2
Dump: lvol2     on:     /dev/disk/disk1_p2, 0

Step 8

Add a new disk name in /stand/bootconf file. Here l means the disk is managed by lvm.

# echo "l /dev/disk/disk2_p2" >> /stand/bootconf

Lastly, set a new disk’s hardware path as an alternative boot path, and you are done. Hardware path of the new disk can be obtained from ioscan output.

# setboot -a 2/0/0/2/0.6.0

Verify it

# /usr/sbin/setboot
Primary bootpath : 2/0/0/3/0.0x6.0x0 (/dev/rdisk/disk1)
HA Alternate bootpath :
Alternate bootpath : 2/0/0/2/0.0x6.0x0 (/dev/rdisk/disk2)
 
Autoboot is ON (enabled)
Hyperthreading : OFF
               : OFF (next boot)

Your root mirror on new disk completed!!

You can reboot the system and try to boot from alternate disk from EFI to test.

Comments

  1. Adolfo Leon Hernandez Abadia says

    Buenos dias,

    En un servidor que estaba realizando los pasos para realizar el mirror, tenia la siguiente.

    1. En el vg_home_users estan montados dos filesystem /home/or11gr24 y /oracle

    2. Cometi el error de ejecutar sudo idisk -wf /tmp/partitionfile /dev/rdisk/disk11 y sudo mkboot -e -l /dev/rdisk/disk11 cuando cai la cuenta era sobre el PV /dev/rdisk/disk3.

    3. En este momento he tratado de solucionar el tema, asi:

    sudo mount -a
    UX:vxfs mount: ERROR: V-3-26881: Cannot be mounted until a full file system check (fsck) is performed on /dev/vg_home/lvol. Please refer to fsck_vxfs man page for details.
    mount: /dev/vg00/lvol1 is already mounted on /stand
    mount: /oracle: I/O error

    bdf
    bdf: /oracle: I/O error

    sudo fsck -F vxfs /dev/vg_home_users/rlv_home_or11gr24
    UX:vxfs fsck: WARNING: V-3-20836: file system had I/O error(s) on meta-data.
    UX:vxfs fsck: ERROR: V-3-26248: could not read from block offset devid/blknum 0/4194296. Device containing meta data may be missing in vset or device too big to be read on a 32 bit system.
    UX:vxfs fsck: ERROR: V-3-20694: cannot initialize aggregate
    file system check failure, aborting …

    sudo cat /etc/fstab |grep vg_home_users
    /dev/vg_home_users/lv_home_or11gr24 /home/or11gr24 vxfs rw,suid,largefiles,delaylog,datainlog 0 2
    /dev/vg_home_users/lvora9 /oracle vxfs rw,suid,largefiles,delaylog,nodatainlog 0 2

    fsck -F vxfs -y -o full /dev/vg_home_users/disk11
    UX:vxfs fsck: ERROR: V-3-20945: cannot stat /dev/vg_home_users/disk11

    sudo vgreduce /dev/vg_home_users /dev/disk/disk11
    vgreduce: Physical volume “/dev/disk/disk11” could not be removed since some of its
    physical extents are still in use.

    sudo vgchange -a n /dev/vg_home_users
    vgchange: Couldn’t deactivate volume group “/dev/vg_home_users”:
    Device busy

    sudo umount /home/or11gr24
    umount: cannot find /home/or11gr24 in /etc/mnttab
    cannot unmount /home/or11gr24

    sudo umount /oracle
    umount: cannot unmount /dev/vg_home_users/lvora9 : Device busy
    umount: return error 1.

    sudo umount -V /home/or11gr24
    umount: cannot find /home/or11gr24 in /etc/mnttab
    cannot unmount /home/or11gr24

    sudo umount -V /oracle
    umount /oracle

    sudo vgreduce /dev/vg_home_users /dev/disk/disk11
    vgreduce: Physical volume “/dev/disk/disk11” could not be removed since some of its
    physical extents are still in use.

    sudo vgreduce -f /dev/vg_home_users /dev/disk/disk11
    Usage: vgreduce
    [-A Autobackup]
    [-l]
    VolumeGroupName PhysicalVolumePath …

    vgreduce
    [-A Autobackup]
    -f VolumeGroupName
    only VolumeGroupName is needed for -f option

    sudo vgchange -a n /dev/vg_home_users
    vgchange: Couldn’t deactivate volume group “/dev/vgxx”: Device busy

    ll /dev/vg_home_users/group
    crw-r—– 1 root sys 128 0x001000 Feb 2 2020 /dev/vg_home_users/group

    sudo vgscan -v
    Unable to scan “Physical volumes” on system
    Unable to scan “Physical volumes” on system

    sudo strings /etc/lvmtab
    /dev/vg00
    /dev/disk/disk99_p2
    /dev/disk/disk3_p2
    /dev/vguser
    tI{Em
    /dev/disk/disk96
    /dev/disk/disk97
    /dev/disk/disk40
    /dev/disk/disk47
    /dev/vgswap
    /dev/disk/disk92
    /dev/disk/disk93
    /dev/disk/disk1
    /dev/disk/disk21

    sudo pvremove /dev/rdisk/disk11
    pvremove: The physical volume “/dev/rdisk/disk11” belongs to volume group “/dev/vg_home_users”.

    fsck -F vxfs -o full /dev/vg_home_users/disk11
    UX:vxfs fsck: ERROR: V-3-20945: cannot stat /dev/vg_home_users/disk11

    sudo rmboot /dev/rdisk/disk11

    lvrmboot -r /dev/vg_home_users
    No valid Boot Data Reserved Areas exist on any of the
    Physical Volumes in the Volume Group “/dev/vg_home_users”.
    Issue the pvcreate -B command to create a Boot Area on a Physical Volume.

    Reply

Share Your Comments & Feedback:

This site uses Akismet to reduce spam. Learn how your comment data is processed.