Scan for Disks

Run the command below to scan for disks.

ioscan -fnC disk

Make your physical volume

This is the first step in the creation of a physical volume.

pvcreate -f /dev/rdsk/c0t1d0

Create the device structure using mknod

Mknon creates a FIFO, character special file, or block special file with the
specified name. You will need to do this when you are creating a volume group.

              cd /dev

              mkdir vgdata

              cd vgdata

              mknod group c 64 0x010000

What is above is only an example, your minor number will vary by machine.

     mknod [options]... NAME Type [Major Minor]

Create volume group

Using vgcreate,you need to create you volume group and list all the disks
that are to make up this volume group. In the example below we are creating a
volume group called vgdata.

                vgcreate vgdata  /dev/dsk/c0t1d0

Below is an example from one of the DCA boxes:

                 vgcreate /dev/vgapcj06 /dev/dsk/c7t2d2 /dev/dsk/c7t2d3 /dev/dsk/c7t2d4

Once created use vgdisplay to display information pertaining to your new
volume.

                 vgdisplay -v vg01

updca003(root)#vgdisplay vgapcj06

--- Volume groups ---

VG Name                     /dev/vgapcj06

VG Write Access             read/write

VG Status                   available

Max LV                      255

Cur LV                      0

Open LV                     0

Max PV                      16

Cur PV                      3

Act PV                      3

Max PE per PV               8690

VGDA                        6

PE Size (Mbytes)            4

Total PE                    26067

Alloc PE                    0

Free PE                     26067

Total PVG                   0

Total Spare PVs             0

Total Spare PVs in use      0

Create a Logical Volume

See the example below that was created with the -l option which used logical
extents to determine the size for the logical volume

updca003(root)#lvcreate -n ndata6 -l 26067 /dev/vgapcj06

Logical volume "/dev/vgapcj06/ndata6" has been successfully created with

character device "/dev/vgapcj06/rndata6".

Logical volume "/dev/vgapcj06/ndata6" has been successfully extended.

Volume Group configuration for /dev/vgapcj06 has been saved in /etc/lvmconf/vgapcj06.conf

updca003(root)#lvdisplay /dev/vgapcj06/ndata6

--- Logical volumes ---

LV Name                     /dev/vgapcj06/ndata6

VG Name                     /dev/vgapcj06

LV Permission               read/write

LV Status                   available/syncd

Mirror copies               0

Consistency Recovery        MWC

Schedule                    parallel

LV Size (Mbytes)            104268

Current LE                  26067

Allocated PE                26067

Stripes                     0

Stripe Size (Kbytes)        0

Bad block                   on

Allocation                  strict

IO Timeout (Seconds)        default

Make a filesystem

   updca003(root)#newfs /dev/vgapcj06/rndata6

   newfs: /etc/default/fs is used for determining the file system type

   version 4 layout

   106770432 sectors, 13346304 blocks of size 8192, log size 256 blocks

   unlimited inodes, largefiles not supported

   13346304 data blocks, 13345512 free data blocks

   408 allocation units of 32768 blocks, 32768 data blocks

   last allocation unit has 9728 data blocks

Other LVM Commands for HPUX

Export disk group

• vgexport /dev/disgroup

this deactivates the diskgroup

• vgchange -a n <diskgroup>

this activates a diskgroup

• vgchange -a y <diskgroup>

Grow a logical volume

• lvextend

Below are instructions on how to modify and confirm your bootlist, which is the list of devices that your AIX system is going to try to boot from. Below is a list of quick and easy commands.

• To specify the logical volume hd5 on disk hdisk0 for normal boot, typle
  • bootlist -m normal hdisk0 blv=hd5

• To view the boot list that you set in the example above
  • bootlist -m normal -o      #where normal sepcifies normal boot

• To specify booting in normal mode from the only boot logical volume on hdisk0, or the boot record hd5 on hdisk1, do the following
  • bootlist -m normal hdisk0 hdisk1 blv=mb_hd5
• If you are dealing with a new disk you are going to need to initialize the boot block as well.
  • In this example you can see that the boot block on hdisk0 has not been initialized.

  • hdisk1 blv=hd5

    hdisk0

  • # bosboot -ad /dev/hdisk0 #use the bosboot command to initialize it.
  • Now all is well

    • # bootlist -m normal -o

      hdisk1 blv=hd5

      hdisk0 blv=hd5

Below are some quick and dirty commands that maybe usefull.

bootlist -m (normal or service) -o –
displays bootlist
bootlist -m (normal or service) (list of devices) – change bootlist
bootinfo -b –
Identifies the bootable disk
bootinfo -t   –
Specifies type of boot
bosboot -a -d (/dev/pv) – Creates a complete
boot image on a physical volume.
mkboot -c -d (/dev/pv)       Zero’s out
the boot records on the physical volume.
savebase -d (/dev/pv)        Saves
customised ODM info onto the boot device.

Need to replace a failed rootdisk in AIX? The post below outlines the procedure for removing a disk from a mirrored root volume.

First take a look at what disks are in your rootvg.

In the example below we have hdisk0 and hdisk1

# lsvg -p rootvg
rootvg:
PV_NAME
PV STATE          TOTAL
PPs   FREE PPs    FREE DISTRIBUTION
hdisk0
active
542         264
109..00..00..46..109
hdisk1
 active
 542         278
 88..00..00..81..109

 lsvg -l rootvg

 LV
 NAME
 TYPE       LPs   PPs
 PVs  LV STATE      MOUNT POINT

 hd5
 boot       1
 2     2    closed/syncd  N/A

hd6

paging     64    128

2    open/syncd    N/A

hd8

jfslog     1     2

2    open/syncd    N/A

hd4

jfs        6

12    2    open/syncd

/

hd2

jfs        127   254

2    open/syncd    /usr

hd9var

jfs        4

8     2    open/syncd

/var

hd3

jfs        37

74    2    open/syncd

/tmp

hd1

jfs        3

6     2    open/syncd

/home

hd10opt

jfs        21

42    2    open/syncd

/opt

dumplv
 sysdump    14    14
 1    open/syncd    N/A

Now lets check to make sure that there’s a 1:2 relationship, meaning that there are copies.

Notice dumplv. It’s not copied so we need to make sure dumplv data isn’t on the failing disk. To check, run;

$ lslv -l dumplv
dumplv:N/A
PV
COPIES        IN BAND
DISTRIBUTION
hdisk0
014:000:000   100%
000:014:000:000:000

This is telling us that the logical volume dumplv is on hdisk0.  If hdisk1 is the failing
disk, then we are okay.  Otherwise, we would have to migrate the data over to the good drive and proceed.

# unmirrorvg rootvg hdisk1
# reducevg rootvg hdisk1
# rmdev -l hdisk1 –d

Before you power down, you need to make sure that your system is set to boot off of a disk that is still in the rootvg.

# bootinfo -b
hdisk1

This tells you what drive it was last booted up.  We want to change this to boot to the new drive, so;

# bosboot -ad
 /dev/hdisk0

And check bootlist

# bootlist –m normal
 –o

NOW WE CAN POWER DOWN THE BOX AND REPLACE THE DRIVE. Once disk has been replaced,
power up the server.  Once at command prompt, run;

# cfgmgr

This will install the new device and allow the OS to see it.

 lsdev -Cc disk

 hdisk0
 Available 40-60-00-4,0 16 Bit LVD SCSI Disk Drive

 hdisk1
 Available 40-60-00-8,0 16 Bit LVD SCSI Disk Drive

Make sure that the OS says it’s available.  If it is, we can assign it to a volume group.

# extendvg rootvg hdisk1

 This will assign it a PVID. and assign it to the volumegroup rootvg to make it available for use.  Now we can mirror;

# mirrovg rootvg

This will take a little while as it’s taking all data now on hdisk0 and making a copy to hdisk1.

# lsvg -p rootvg
rootvg:
PV_NAME
PV STATE          TOTAL
PPs   FREE PPs    FREE DISTRIBUTION
hdisk0
active
542         264
109..00..00..46..109
hdisk1
 active
 542         278
 88..00..00..81..109

Once it’s mirroring, we can make sure it’s assigned to rootvg by doing the above.  We can also check to make sure there’s copies;

# lsvg -l rootvg
rootvg:
LV
NAME
TYPE       LPs   PPs
PVs  LV STATE      MOUNT POINT
hd5
boot       1
2     2    closed/syncd  N/A
hd6
paging     64    128
2    open/syncd    N/A
hd8
jfslog     1     2
2    open/syncd    N/A
hd4
jfs        6
12    2    open/syncd
/
hd2
jfs        127   254
2    open/syncd    /usr
hd9var
jfs        4
8     2    open/syncd
/var
hd3
jfs
37    74    2
open/syncd    /tmp
hd1
jfs        3
6     2    open/syncd
/home
hd10opt
jfs        21
42    2    open/syncd
/opt
dumplv
 sysdump    14    14
 1    open/syncd    N/A

Now we need to modify the bosboot to recreate the boot image.

# bosboot –a

Double check your bootlist to make sure hdisk1 is in there.

# bootlist –m normal –o