hpux lvm new features

5/28/2018 HPUX Lvm New Features

1/35

1

11i v3LVM New Features in HP-UX

Abstract ........................................................................................................................................................................................

........................

........................

........................

........................

........................

........................

34444889

........................9

........................9

......................10

......................10......................14

......................14

......................15

......................16

......................16

......................16

......................17

......................17

......................17

......................18

......................18

......................19

......................19

......................19

......................20

......................20pvdisplay Displays user data offset and checks if disk is under HP-UX LVM control ......................................21vgscan - Scans Faster, Per Volume Group Scanning and Supports Persistent DSF ..........................................22vgcfgrestore Enhanced Listing of Backup File Contents ................................................................................25Error Management Technology (EMT)..............................................................................................................26Long Host Name Support for Display Commands .............................................................................................26lvmadm Command ...........................................................................................................................................27vgremove Removal of Volume Group Device Files ........................................................................................ 28vgdisplay Displays More Information ............................................................................................................28

New functionality ........................................................................................................................

of Mass StorageSupport for Agile View .....................................................................................

Multipathing and Alternate Links (PVlinks) .................................................................................

Dynamic LUN Expansion (DLE) ..................................................................................................Modification of Volume Group Settings ......................................................................................

Modification of Physical Volume Type (Boot/Non-Boot) .............................................................SLVM Online Reconfiguration .....................................................................................................LVM Device Online Replacement (OLR)......................................................................................

Volume Group Quiesce and Resume ..........................................................................................Boot Resiliency ...........................................................................................................................

Striped and Mirrored Logical Volumes ........................................................................................Better Coexistence with Other Disk Users ..................................................................................Better Utilization of Disk Space No Bad Block Reserve Area...................................................Elimination of maxvgs Tunable ................................................................................................ ..

Proactive Polling of Alternate links .............................................................................................Version 2.x Volume Groups ........................................................................................................

Boot Disk Configuration Self Healing..........................................................................................

Performance Improvements ......................................................................................................

Mirror Write Cache (MWC) Enhancements .................................................................................LVM Support for Large I/O Sizes ................................................................................................

Parallel Resynchronization of Stale Extents ................................................................................Parallel Open of Physical Volumes..............................................................................................

Increased Limits ..........................................................................................................................

Maximum Logical Volume Size Increased up to 16TB ................................................................Increased Limits for Version 2.x Volume Groups ........................................................................

Usability Enhancements .............................................................................................................

Compact, Parsable Command Output ........................................................................................


2/35

2

......................30

......................30

......................32

......................32

......................32

......................33

......................35Call to Action......................................................................................................................................................35

Auto Creation of Volume Group File ...........................................................................................pvmove Enhancements ..............................................................................................................

Miscellaneous...............................................................................................................................

Commands enhanced to prevent misconfigurations through alternate links ..............................Mirror Disk Installation No Longer Requires a Reboot................................................................

Glossary ........................................................................................................................................

For More Information .................................................................................................................


3/35

3

Abstract

d availability007, there have been three

updates of HP-UX 11i v3:September 2007, March 2008, and September 2008. This whitepaper lists

Some features have their own whitepaper; this document only summarizes them. For more information, see the

In HP-UX 11i v3 (11 .31), LVM delivers significant performance, scalability, usability, anenhancements. After the initial release of HP-UX 11i v3 in February 2

all the new LVM features in the initial release and the three updates of HP-UX 11i v3.

referenced documents in the For More Information section. Other features are presented in moincluding their use and

re detail,benefits.

The document is intended for system administrators, operators, and customers who want to use and know about newLVM features in HP-UX 11i v3.


4/35

4

y

le view. In thisnk) information. The

s to it. The legacycontinues to exist. You can use both the DSF types to access a given

r more information, see the The Next

New functionalit

Support for Agile View of Mass Storage

HP-UX 11i v3 introduces a new representation of mass storage devices called the agirepresentation, the device special file (DSF) name for each disk no longer contains path (or li

multipathed disk has a single persistent DSF regardless of the number of physical path

view, represented by the legacy DSF,mass storage device independently and both can coexist on a system. FoGeneration Mass Storage Stackwhitepaper.

configuration commands support both DSF naming models.

ombination of

HP recommends the use of persistent DSFs for LVM configurations and encourages configuring new volumee all the capabilities of the new mass storage stack, HP recommends

migrating volume groups from legacy DSFs to persistent DSFs. HP provides/usr /contrib/bin/vgdsftoore information, see the

Wherever applicable, LVM

LVM enables you to configure volume groups with all persistent DSFs, all legacy DSFs, or a c

persistent DSFs and legacy DSFs.

groups using persistent DSFs. To fully utiliz

facilitate this migration. The script works for both root and non-root volume groups. For m

LVM Migration from Legacy to Agile Naming Modelwhitepaper.

Multipathing and Alternate Links (PVlinks)

mass storage stack.istent DSF for each mass storage device regardless of the number of

e transparent

Management of multipathed devices is available outside of LVM using the next generationAgile addressing creates a single pershardware paths to the disk. The mass storage stack in HP-UX 11i v3 uses that agility to providmultipathing. For more information, see the The Next Generation Mass Storage Stackwhitepaper.

ty is now redundant, but this functionality is still supported with legacy

they did in previous releases when the mass storage stack native multipathingfeature is disabled with the scsimgrcommand.

the use of persistent

sfscript, vgscan-Ncommand, or vgimport-s -Ncommands perform thisconversion. For more information, see the LVM Migration from Legacy to Agile Naming Model

The LVM Alternate Link functionali

DSFs. Alternate links behave as

HP recommends converting volume groups with alternate links to use native multipathing by

DSFs. The/usr/contrib/bin/vgd

whitepaper.

Some disk arrays allow the dynamic resizing of their LUNs. With HP-UX 11i v3, LVM detects and handles physical

for Version 1.0

VM supports, see the

Dynamic LUN Expansion (DLE)

volume size changes on invocation of the vgmodifycommand.This command is supported

volume groups only. For more information on different versions of volume groups that LVersion 2.x Volume Groups section.

vgmodify(1M)

The vgmodifycommand provides a method to alter the attributes of a physical volume and volume group after

pvcreateand vgcreateoperations, respectively. You must run the vgmodifycommand to update the LVM configurationto reflect any change to the physical volume size.

Prior to September 2008 release of HP-UX 11i v3, you can alter the attributes of a physical volume or a volume

group using the vgmodifycommand only when the volume group is deactivated. Starting with the September

2008 release of HP-UX 11i v3, LVM supports handling dynamic LUN expansion of physical volumes using the

vgmodifycommand when the volume group is active. This is referred to as online DLE. To support online DLE,
http://docs.hp.com/en/MassStorageStack/The_Next_Generation_Mass_Storage_Stack.pdfhttp://docs.hp.com/en/LVMmigration1/LVM_Migration_to_Agile.pdfhttp://docs.hp.com/en/LVMmigration1/LVM_Migration_to_Agile.pdfhttp://docs.hp.com/en/MassStorageStack/The_Next_Generation_Mass_Storage_Stack.pdf


5/35

5

LVM introduces two new options: -a(to indicate an online operation) and -E(to handle a DLE operation

information, see vgmodify(1M) and the - LVM Volume Group Dynamic LUN expansion (DLE)/vgmodify

explicitly).

For more

as enhanced in the February 2007 release of HP-UX 11i v3 to allow relocation of just the

pand the LVMs on-

the existing allocation policy of a logical volume containing the extent that is

considered for relocation.

For more information, see pvmove(1M)

t to any free extent within the same physical volume, enter:

t to any free extent in the volume group, enter:

:0

a physical volume that vdisplay(1M). You can then relocate the first user extent to

sical volume using p

/ dev/ di sk/ di sk22/ dev/ vgnameavai l abl e

VGDA 2

4

meout ( Seconds) def aul tAut oswi t ch On

ation physical volume

# pvmove /dev/disk/disk10:0 /dev/disk/disk22

Note: Relocation of the first data extent fails if it violates the strict mirror allocation policy. So identify asuitable physical volume that does not hold any mirror copies of the same extent which is considered forrelocation.

To relocate the first data extent of a physical volume when the vgmodify commandreports insufficient spacefor expanding the LVM configuration data on the disk. This happens when the user tries to modify the

physical volume settings using the vgmodify command.

whitepaper.

pvmove(1M)

The pvmove command w

first extent of the physical volume. The vgmodify command uses this additional space to exdisk configuration information.

The pvmove command honors

Examples

To relocate the first data exten

# pvmove /dev/disk/disk10:0 /dev/disk/disk10

To relocate the first data exten

# 0d0pvmove /dev/dsk/c1t

To find has free space, use p

that phy vmove:

# pv /d display /dev/disk isk22

- - - Physi cal vol umes - - -PV Name

eVG NamPV Stat usAl l ocat abl e yes

Cur LV 5PE Si ze ( Mbyt es)

Tot al PE 1279Free PE 779

Al l ocated PE 500St al e PE 0I O Ti

To relocate first data extent from the source physical volume to the desired destin

that was identified in the last step, enter:


6/35

6

as follows:

abase

roups - - -

v/ vgdat abasee Access

at us avai l abl eV 255V 1

1

16

1per PV 2559

2( Mbyt es) 4

2559102549

TT i n use 0

al vol umes -ame / dev/ vgt est / l vol 1

LV St at us avai l abl e/ syncd

t LEed PE 0

V 1

at us avai l abl eTot al PE 2559

2549

On

The vgmodify command can report up s to adjust the number of extents and physical

volume upwards, where possible spac rved on each physical volume for the LVM

et t i ngs:

V 2559Mb ) 4

) 400" ar e i ncr eased f r om2559 t o 6652

The cur r ent and new Vol ume Gr oup paramet er s di f f er .An updat e t o t he Vol ume Gr oup I S r equi r ed

New Vol ume Gr oup set t i ngs:Max LV 255Max PV 16

Max PE per PV 6652PE Si ze ( Mbyt es) 4

VGRA Si ze ( Kbyt es) 896Revi ew compl ete. Vol ume group not modi f i ed

Consider a volume group with on

# vgdisplay -v /dev/vgdat

e physical volume and one logical volume

- - - Vol ume g

eVG Name / dr ead/ wr i t eVG Wr i t

VG StMax LCur LOpen LV

Max PVCur PV 1Act PVMax PEVGDAPE Si ze

Tot al PEAl l oc PEFr ee PE

GTot al PV 0ot al Spar e P 0Vs

e PVsotal Spar

- - - -LV N

Logi c

LV Si ze ( Mbyt es) 400Cur r en 1

t 1Al l ocaPUsed

- - Physi cal vol umes - - -PV Name / dev/ dsk/ c11t0d5PV St

Fr ee PE

Aut oswi t ch

optimized volume gro ingsett

, to make full use of the e reseconfiguration data:

# vgmodify -o -r /dev/vgtest

Curr ent Vol ume Gr oup s Max LV 255

Max PV 16Max PE per P

PE Si ze ( yt es

VGRA Si ze ( Kbyt esNew conf i gur at i on requi r es "max_pes


7/35

7

value of Max PVfiguration data), it

ifycommand to check

atively, you can relocatethe Max PV value without

mmand.

aximum physical volume setting in the volume group to a value greater than 16

st

Curr ent Vol ume Gr oup setLV 255PV 16

Max PE per PV 2559

VGRA Si ze ( Kbytes) 400onl y be completed i f PE number zero on

f r eed

st data extent. Use the pvmove command to

s follows:

up

0d5:0

me "/ dev/ vgt est / l vol 1". . .uccessf ul l y moved.

n f or / dev/ vgt est has been saved i n

ccessf ul l y changed.

PE V 2559) 4

400r .

New Vol ume Gr oup set t i ngs:Max LV 255Max PV 64

Max PE per PV 2559PE Si ze ( Mbyt es) 4

VGRA Si ze ( Kbytes) 1488New Vol ume Gr oup conf i gur at i on f or "vgt est " has been saved i n"/ et c/ l vmconf / vgt est . conf "Ol d Vol ume Gr oup conf i gur at i on f or "vgt est " has been saved i n"/ et c/ l vmconf / vgt est . conf . ol d"

The previous output shows that you can change the Max PE per PV value to 6652 with theretained at 16. To increase the value of Max PV (with the current available space for LVM con

might be necessary to reduce the value of Max PE per PV. Use the toption of the vgmod

if the Max PV value can be increased by changing the value of Max PE per PV. Alternthe first data extent to create more space for LVM configuration data and increase

reducing the Max PE per PV value. To achieve this, use the noption of the vgmodifyco

Any attempt to increase the mresults in a failure as follows:

# vgchange -a n /dev/vgte

Vol ume gr oup "vgtest " has been successf ul l y changed.

# vgmodify -n -p 64 /dev/vgtest

t

i ngs:Max

Max

PE Si ze ( Mbyt es) 4

vgmodi fy: Thi s operati on can"/ dev/ r dsk/ c11t 0d5" i s

Note that the last output message indicates you must free the fir

relocate the first data extent a

# vgchange -a y /dev/vgtest

Act i vat ed vol ume gr oVol ume gr oup " / dev/ vgt est " has been successf ul l y changed.

# pvmove /dev/dsk/c11t

Transf err i ng l ogi cal extent s of l ogi cal vol uk/ c11t 0d5" has been sPhysi cal vol ume "/ dev/ ds

Vol ume Gr oup conf i gurat i o/ et c/ l vmconf / vgt est . conf

Use the vgmodify command to increa ph volume setting. This time it succeeds:se the maximum ysical

# vgchange -a n /dev/vgtest

Vol ume gr oup "vgtest " has been su

# vgmodify -n -p 64 /dev/vgtest

Cur r ent Vol ume Gr oup set t i ngs:Max LV 255Max PV 16

Max r PpePE Si ze byt e ( M s

VGRA Si ze ( Kbytes)The cur r ent and new Vol ume Gr oup paramet ers di f f e

An updat e t o t he Vol ume Gr oup I S r equi r ed


8/35

8

St ar t i ng t he modi f i cat i on by wr i t i ng t o al l Physi cal Vol umesAppl yi ng the conf i gur at i on to al l Physi cal Vol umes f r om "/ et c/ l vmconf / vgt est . conf "

has been saved i n

Now issuing vgdisplayon the volume group shows the modified values (in this case the maximum physical volume)follows:

est

"vgt est " uccessf ul l y changed.

e / dev/ vgt estt e Access r ead/ wr i t e

at us avai l abl e255

64

11V 2559

24255810

Fr ee PE 2548

Modification of Volume Group Settings

max_pe, max_pv,

on an existing

ing three configuration parameter set via vgcreate:

ax_pe setting set by

vgcreate -e).

lume group can contain (max_pv setting set by vgcreate-

p).

tting set by vgcreate -l).

The vgmodifycommand displays the possible volume group max_pe and max_pv settings for this volume group tohelp optimize an existing volume group configuration.

For more information, see vgmodify(1M) and the LVM Volume Group Dynamic LUN expansion (DLE)/vgmodify

Compl et ed t he modi f i cat i on pr ocess.New Vol ume Gr oup conf i gur at i on f or "vgt est ""/ et c/ l vmconf / vgt est . conf . ol d"Vol ume gr oup "vgtest " has been successf ul l y changed.

for the volume group as

# vgchange -a y vgt

Act i vated vol ume gr oupsVol ume group has been

# vgdisplay /dev/vgtest

VG Namr iVG W

VG StMax LVCur LV 1

1Open LVMax PV

Cur PVAct PVMax PE per PVGDAPE Si ze ( Mbyt es)

Tot al PEAl l oc PE

Tot al PVG 0Tot al Spar e PVs 0Tot al Spar e PVs i n use 0

When an LVM volume group is created, several configuration parameters are set (for example,

max_lv). The new vgmodifycommand enables you to change these configuration parametersvolume group, which avoids having to migrate user data.

The vgmodifycommand can alter the follow

The maximum number of physical extents that can be allocated per physical volume (m

The maximum number of physical volumes that the vo

The maximum number of logical volumes that the volume group can contain (max_lv se

whitepaper.

Modification of Physical Volume Type (Boot/Non-Boot)

When initializing a physical volume for LVM, pvcreateassigns a type to it, either boot or non-boot (Note thatbootable physical volumes are not supported in Version 2.xvolume groups).


9/35

9

boot or vice versa.

pvcreate(1M) for the -Boption. Also see the LVM Volume Group Dynamic LUN expansion

The vgmodifycommand enables you to change a physical volume type from boot to non-

See vgmodify(1M) and

(DLE)/vgmodifywhitepaper.

g a physical volume non-bootable increases the space available on that device for LVMconfiguration data. However, to take advantage of the additional space, you must mark every disk in a volume

Starting with the HP-UX 11i v3 September 2009 release, you can change the configuration of an active shareds stay available on all

nge -xoption, the SLVM SNOR feature to change

the configuration of a shared volume group, while keeping it active only on a single node. Using this procedure,in available during the volume group reconfiguration.

(OLR) feature provides the following new methods for replacing or isolating path

s to a physical

replacing disks in an activeire deactivating the volume group, modifying the volume group

gnostics against them.

sing pvchange -a Ncommand, in order toperform an online disk replacement, is still supported. The behavior is the same for both legacy and

ith previous releases.

roup, the pvchange- a

, use the scsimgrsical volume paths. For more information, see scsimgr(1M).

Note that makin

group non-bootable.

SLVM Online Reconfiguration

volume group in a cluster without deactivating the volume group on any node. Applicationnodes sharing the volume group. This applies to Version 2.1 and higher volume groups.

For Version 1.0 and 2.0 volume groups, you can use the vgcha

applications on at least one node rema

For more information, see the SLVM Online Reconfiguration whitepaper.

LVM Device Online Replacement (OLR)

The LVM Online Disk Replacementcomponents or LVM disks within an active volume group:

Using -nand -Noptions with pvchangecommand, you can detach a specific path or all pathvolume, respectively. LVM OLR enables you to follow a simpler procedure forvolume group. The procedure does not requconfiguration, or moving any user data.

Using LVM OLR, you can easily isolate troublesome paths or disks to allow running dia

In HP-UX 11i v3, the option of detaching an entire physical volume u

persistent DSFs and is compatible w

Unless native multipathing is disabled and only legacy DSFs are configured for the volume g

n command does not stop I/O operations for that path as they did in earlier releases. Insteadcommand with the disable option to disable phy

For more information, see the LVM Online Disk Replacement (LVM OLR)whitepaper.

Volume Group Quiesce and Resume

The LVM volume group Quiesce/Resume feature allows quiescing I/O operations to the disks in a volume groupto facilitate creating a consistent snapshot of an otherwise active volume group for backup purposes. This featureworks with backup management and disk array management software to enable them to create a consistentsnapshot of the disks that make up an LVM volume group.

The Quiesce/Resume feature prevents the disk images from changing and allows a snapshot of the disks to betaken without having to unmount or close the open logical volumes and deactivate the volume group.

The vgchange command provides new options -Qand -Rto allow quiescing the volume group prior to creating a

snapshot and to resume the volume group afterward. Optionally, both reads and writes or just writes to thevolume group can be quiesced.


10/35

10

For more information, see theLVM Volume Group Quiesce/Resumewhitepaper.

Boot Resiliency

Root volume group scanning is a new LVM feature in HP-UX 11i v3. The feature can prevent boot failures that can

n information is incorrect or out-

s are: The root volume group is configured using legacy DSFs representing the devices in a Storage Area

ch situations; LVM now scans all thedisks to identify the ones belonging to the root volume group and retries activation. If the activation succeeds, it is

ut of sync with theriver prints a warning:

LV act i vat i on r equi r ed a scan. The PV i nf or mat i on i n t he on- di sk

BDRA f rom t he syst em' s cur r ent I O conf i gur at i on. To updat e t heat e the on- di sk

r oot VG name i s / dev/ vg00:

ot volume group are not available but the quorum is met, no root volume

group scan is performed. Also, during a single user mode boot with -isor maintenance mode boot with -lm, root

Note: This feature is enhanced further in the September 2008 release of HP-UX 11i v3. For more information,nfiguration Self Healing

occur on prior HP-UX releases.

During boot, the root volume group activation can fail if the LVM boot configuratio

auseof-date with the systems current I/O configuration. Two possible c

Network(SAN) and the SAN is reconfigured such that DSFs of the devices changed.

The root disk is relocated to a different slot such that the DSF name changes.

With the new root volume group scanning, LVM automatically handles su

likely that the LVM in-memory boot configuration information for the root volume group is oDSF in the /etc/lvmtab for the root volume group. To assist recovery in this case, the LVM dmessage to the console and logs into /var/adm/syslog/syslog.log to the effect as follows

M: WARNI NG: Root VG

may be out- of - dat eon- di sk BDRA, f i r st update / et c/ l vmt ab usi ng vgscan( 1M) , t hen updBDRA usi ng l vl nboot ( 1M) .For exampl e, i f t he

vgscan - k - f / dev/ vg00l vl nboot - R / dev/ vg00

In case some physical volumes in the ro

volume group scanning is skipped.

see the Boot Disk Co section.

an thet possible stripe size).

monly referred to as striping, refers to the segmentation of logical sequences of data across disks. RAID1, commonly referred to as mirroring, refers to creating exact copies of logical sequences of data. When

implemented in a device hardware, RAID 10 (or RAID 1+0) and RAID 01 (or RAID 0+1) are nested RAID levels. The

difference between RAID 0+1 and RAID 1+0 is the location of each RAID system RAID 0+1 is a mirror ofstripes whereas RAID 1+0 is a stripe of mirrors. Figure 1 shows the RAID 10 and RAID 01 configurations (A1, A2...Axare stripe chunks of a logical volume).

With a hardware-based RAID 10 configuration, I/O operations are striped first then each strip is mirrored. Withhardware-based RAID 01, I/Os are mirrored first then striped. RAID 10 and RAID 01 can have the same physicaldisk layout.

Striped and Mirrored Logical Volumes

HP-UX LVM now introduces support for striped and mirrored logical volumes at a smaller granularity thextent size (4KB is the smalles

RAID 1+0 and RAID 0+1

RAID 0, com


11/35

Figure 1: RAID 1+0 and RAID 0+1

The advantages of hardware-based RAID 10 over RAID 01:

11

When one disk fails and is replaced, only the amount of data on this disk needs to be copied or re-

before data becomes unavailable.

The advantages of hardware-based RAID 01 over RAID 10:

ecause it does not

ut processes them simultaneously. Typically with hardware solutions, you create a LUN witha RAID level and the RAID functions are stacked. LVM provides more flexibility on how logical volumes are

stripe width multiplied by thed striped across two disks,extents of a single set areirrored to obtain the data

irroring in LVM combines the advantages of the hardware implementation of RAID 1+0 and RAID0+1, and provides the following benefits:

Better write performance. Write operations take place in parallel and each physical write operation isdirected to a different physical volume.

Excellent performance for read. Even in the case where several disks are out of service, the read of a stripe

can be done in parallel on different physical volumes with one I/O operation per physical volume.

High availability of data. With multiple copies of the user data residing on different physical volumes, LVMavoids single point of failure ensuring high availability.

Configuration

synchronized.

RAID 10 is more tolerant to multiple disk failures

Simpler to configure striped volumes and then extend mirroring.

Able to split the mirror copy and have two usable volume sets.

LVM Implementation of RAID levels in HP-UX

LVM implementation of RAID management is different from the hardware based solutions b

nest the RAID levels, b

created amongst a set of disks as compared to hardware solutions.

LVM allocates the physical extents for striped and mirrored logical volumes in sets ofnumber of copies of the data. For example, if the logical volume is 1-way mirrored anextents are allocated to the logical volume, four at a time. LVM enforces that the physicalfrom different physical volumes. Within this set, the logical extents are stripped and mlayout displayed in Figure 1.

Striping and m


12/35

12

irror to) an existinging can be directed to

are based RAIDg the striping and mirroring processing, which allows LVM to provide the best of both

RAID 10 and RAID 01.

size and stripe width

you to create airrored logical volume as long as there are enough free extents in the volume group. The logicalvolu layout then uses a slight variation of RAID 10 and all the benefits of RAID 10 are retained. For example,Figure 2 shows a volume group with seven physical volumes and a logical volume striped across three disks withone mirror copy.

RAID 10, only sixal volumes, PV0 does

ation. PV0 was shorttwo extents were allocated from PV1. LVM created the logical volume using seven physical

volumes resulting in a slight variation of RAID 10.

You can extend your existing striped logical volumes to a striped and mirrored configuration easily, using the

command to add mirrors. You cannot convert existing mirrored logical volumes to a striped and mirrored

configuration as explained in

ing and merging of striped and mirrored logical volumes. These logical volumes are alsos d in exclusive mode activation.

E

d across two physical volumes with one mirror copy and stripe

# lvcreate -L 90 -i 2 -I 64 -m 1 /dev/vgtest

Note: Striping with mirroring always uses strict allocation policies where copies of data do not exist onthe same physical disk. This results in a configuration similar to the RAID 01 as illustrated in Figure1.

To create a logical volume of size 90MB striped across two physical volumes with one mirror copy and stripesize of 64 KB and to create the mirror copies on specific disks ( configuration equivalent to RAID 01 asillustrated in Figure1, enter)

# lvcreate -L 90 -i 2 -I 64 /dev/vgtest

# lvextend -m 1 /dev/vgtest/lvol1 /dev/dsk/c3t0d0 /dev/dsk/c3t1d0

You can create a new striped and mirrored logical volume from scratch, or extend (add mstriped logical volume. With the use of physical volume groups (PVGs), mirroring and stripspecific physical volumes. Without the constraint created by the nesting of levels in hardwsolutions, LVM is combinin

To create a pure RAID 10, ensure that the physical volumes have enough extents in terms of

to accommodate the logical volume. Otherwise, LVM optimizes the extent allocation to enablestriped and mme

Figure 2: Example of LVMs implementation of RAID 10 variation.

First mirror

PV0 PV2 PV3

PV0 PV2 PV3

Second

PV1 PV2 PV3

PV1 PV2 PV3

mirror

PV4 PV5 PV6

PV4 PV5 PV6

PV4 PV5 PV6

PV6

Stripes from PV

PV4 PV5

If the first physical volume used in the creation has enough extents (size of LV/3) to createphysical volumes are used, resulting in a strict RAID 1 0 configuration. But if one of the physic

not have enough free extents; LVM uses another physical volume, PV1, to complete the operby two extents so

lvextend

configuration. You can create any new logical volumes with striped and mirroredthe Examples section.

LVM allows the splittupported in a clustered environment with shared an

xamples

Use any of the following procedures to create a striped and mirrored logical volume:

To create a logical volume of size 90MB stripesize of 64 KB, enter.


13/35

13

mpvgtest

0d0

1d0

e has the striped property as shown by the lvdisplayoutput:

l vol umes -/ dev/ vgt est / l vol 1

r ead/ wr i t eavai l abl e/ syncd

striped1024256

64

on

I O Ti meout ( Seconds) def aul t

onfiguration, extend the logical volume as follows:

v/vgtest/lvol1

e volume group physical volumes and extents to accommodate the mirror copy.

put sho e logical volume is striped and mirrored:

dev/vgtes l1

l vol 1/ dev/ vgt estr ead/ wr i t eavai l abl e/ syncd

Mirror copies 1

Consi st ency Recover y MWCSchedule striped

LV Si ze ( Mbyt es) 1024Curr ent LE 256Al l ocat ed PE 256Stripes 2

Stripe Size (Kbytes) 64

Bad bl ock onAl l ocat i on s t r i ctI O Ti meout ( Seconds) def aul t

Contents of /etc/lvVG /dev/vgPVG PVG0

/dev/dsk/c2t/dev/dsk/c2t

PVG PVG1/dev/dsk/c3t0d0/dev/dsk/c3t1d0

Adding a mirror to existing striped logical volumes.

The existing logical volum

# lvdisplay /dev/vgtest/lvol1

- - Logi ca - -LV NameVG Name / dev/ vgt estLV Per mi ssi onLV St atus

Mirror copies 0

over y MWCConsi st ency ReceSchedul

LV Si ze ( Mbyt es)Curr ent LEAl l ocated PE 256Stripes 2

Stripe Size (Kbytes)

Bad bl ockAl l ocat i on s t r i ct

To get a striped and mirrored c

# lvextend -m 1 /de

Note that th must have enough

Now the lvdisplayout ws th

# lvdisplay / t/lvo

- - Logi cal vol umes - - -/ dev/ vgt est /LV Name

VG NameLV Per mi ss i onLV St atus


14/35

14

ot support combinationg and striping is

ed on HP-UX 11i v3, attempts to import or activate its associated volume group fails on a previous HP-UX

To import the volume group with striped and mirrored logical volume to releases prior to HP-UX 11i v3, you must

iped and mirrored logical volume is supported on HP-UX 11i v2 with the PHCO_36744,PHKL_36745tches or superseding patches installed on a HP-UX 11i v2 system,

volume group with striped and mirrored logical volume created onHP-UX 11i v3.

Other Disk Users

s. The LVM commands pvcreateand vgcfgrestorenow check if their target

nmark a LVM disk. This is facilitated by the pvremovecommand. For moreion, see the pvremove(1M) man page.

the disk, overwriting any other

scripts or during initializing a

command.

isk If the disk /dev/dsk/c3t1d1 is marked for use by Oracle ASM, an

ngs to some other subsystem as follows:

pvcr eat e: Coul d not per f or m LVM oper at i on on Or acl e ASM di sk "/ dev/ r dsk/ c3t 1d1"

ecommand as follows:

sk/disk26

removed.

Software bad block relocation refers to capability of LVM to handle I/O errors on disk blocks that are bad. Inprevious releases of HP-UX, when a media failure occurs (detection of a bad block of data on disk), LVM marks

the failed block in the Bad Block Directory and attempts to relocate the block to a new location in the Bad BlockReserve Area(BBRA) on the disk.

The BBRA was created when the disk was initialized using the pvcreatecommand.

All modern disks now offer hardware based bad sector relocation and do not trigger the LVM software relocation.With HP-UX 11i v3, LVM does not reserve a BBRA at the end of disk to allocate more space for user data.

Compatibility Note

Releases prior to HP-UX 11i v3 only support striped or mirrored logical volumes and do nof striped and mirrored logical volumes. If a logical volume using simultaneous mirrorincreatrelease.

remove the incompatible logical volumes or reduce them to no mirrors.

NOTE: Strand PHCO_36746 patches. With these payou can successfully import and activate a

Better Coexistence with

LVM coexists better with other disk user

device is in use by other subsystems.

LVM has a provision to disown or uinformat

Note: pvcreatewith the force option -foverrides this check and takes ownership ofsubsystems data. HP recommends NOT using the -foption by default in any of yourphysical volume using pvcreate

Example

Coexistence with Oracle ASM d

initialization operation on this disk identifies that the disk belo

# pvcreate /dev/rdsk/c3t1d1

To disown or unmark an LVM disk that is not part of any volume group, use the pvremov

# pvremove /dev/rdi

The physical volume associated with "/dev/rdisk/disk26" has been

Better Utilization of Disk Space No Bad Block Reserve Area


15/35

15

block directory. Forhas bad blocks relocated

a HP-UX 11i v3 release, LVM continues to support thise relocated blocks from the BBRA.

256, which could bee system. With HP-UX 11i v3 release, the maxvgs tunable was removed. Users can

now create up to 256 Version 1.0 volume groups and 2048 Version 2.xvolume groups (the combined number of

g setting this kernel

M supports, see the Version 2.x Volume

Compatibility Note

In HP-UX 11i v3, LVM is compatible with physical volumes that already have entries in the badexample, a disk device is configured for LVM on HP-UX 11i v2 and this physical volumeto the BBRA. On importing such a physical volume ontophysical volume and can read th

Elimination of maxvgs Tunable

This static tunable specified the maximum number of LVM volume groups in the range 1 tocreated or activated on th

Version 2.0 and Version 2.1 volume groups cannot exceed 2048) in a system without requirintunable, thus avoiding a reboot.

For more information on different versions of volume groups that LVGroups section and the LVM Version 2.0 Volume Groups in HP-UX 11i v3 whitepaper.

ge scripts that use or modify maxvgs tunable for HP-UX 11i v3.

g

:

# kct

NOTE f ol l owi ng change( s)l l be hel d f or

oma conf i gur at i on cur r ent l y cont ai ns t hei as i n use bef or e t he l ast reboot of t hi s

r ent conf i gurat i onr e maki ng t he r equest ed change? y

i gurat i on has been updat ed.hanges have been saved, and wi l l t ake ef f ect at

Tunabl e Val ue Expr essi onmaxvgs ( now) 10 Def aul t

(next boot) 64 64

On a HP-UX 11i v3 system, no reboot is required as the maxvgs tunable is removed:

# kctune maxvgs

ERROR: There i s no t unabl e named ' maxvgs' .

For more information, see maxvgs(5).

Note: You must chan

Example

On a HP-UX 11i v2 system:

# kctune maxvgs

Tunabl e Val ue Expr essi onmaxv s 10 Def aul t

Modifying the maxvgs tunable on a HP-UX 11i v2 release results in a reboot

une maxvgs=64

: The conf i gur at i on bei ng l oaded cont ai ns t het hat cannot be appl i ed i mmedi at el y and whi ch wi

t he next boot :- - The t unabl e maxvgs cannot be changed i n a dynami c f ashi on.

WARNI N ackup'G: The aut t i c ' bconf i gurat on t hat wsyst em.

==> Do you wi sh to update i t t o contai n t he curbef o

* The aut omat i c ' backup' conf* The request ed c

next boot .


16/35

16

e links

line option in the pvchange

command, which controls the way LVM monitors alternate links to a multipathed physical volume.

By default, LVM proactively monitors the health of all configured links. If auto-switchback is enabled (by

If proactive polling of alternate links is undesirable, you can disable polling using pvchange p n. For more1M).

p- Version 2.0.

at of Version 1.0ps. Version 2.0 enables the configuration of larger volume groups (up to 512), logical volumes (up

to 511), physical volumes (up to 511), and several other volume group parameters. You can manage a Versionace. But featuresupported on Version

ted on Version 2.0

e group- version 2.1.

p, using the sameuser interface. Version 2.1 is similar to version 2.0, but allows configuration of much greater number of volume

s (up to 2048), and logical volumes (up to 2047). With this release, clusterlock is supported on Version 2.xvolume groups.

ps in HP-UX 11i v3

Proactive Polling of Alternat

With September 2007 release of HP-UX 11i v3, LVM supports a new -pcommand

default), this polling helps LVM detect and switch to a better link when it becomes available.

information, see pvchange(

Version 2.xVolume Groups

With the March 2008 release of HP-UX 11i v3, LVM supports a new version of volume grouVersion 1.0 is the version supported on all current and previous versions of HP-UX 11i.

A Version 2.0 volume group is a volume group whose metadata layout is different from thvolume grou

2.0 volume group the same way as a Version 1.0 volume group, using the same user interfsuch as sparing, bad block relocation, cluster lock, boot, dump, and primary swap are not s

2.0 volume groups. Also, vgmodify, pvck, lvlnboot, and lvrmbootcommands are not suppor

volume groups.

With the September 2008 release of HP-UX 11i v3, LVM supports another version of volumYou can also manage a Version 2.1 volume group the same way as a Version 1.0 volume grou

groups (up to 2048), physical volume

For more information on Version 2.x volume groups, see the LVM Version 2.0 Volume Grou xvolume groups, seewhitepaper. For more information on the supported limits of Version 1.0 and Version 2.

the LVM Limits White Paper.

Boot Disk Configuration Self Healing

With the September 2008 release of HP-UX 11i v3, LVM supports self healing of boot disk configurations. If thedevice special file for a LVM boot disk changes because its hardware path changed, LVM scans for the boot diskduring the boot process. After successfully booting, LVM automatically updates the /etc/lvmtab LVMconfiguration file and the LVM metadata on each bootable disk in the root volume group with the new bootpath. Future system boots will not need to scan. This behavior is configurable, and is enabled by default.

For more information, see lvmadm(1M) and lvm(7).


17/35

17

Mirror Write Cache (MWC) Enhancements

ly records thisactivity in an on-disk data structure. An extra disk write is required for every mirrored write not already recorded

WC to resynchronize

logical volume I/O

performance by allowing more concurrent writes. MWC has also been enhanced to support large I/O sizes.

nd, respectively. For

ange(1M).

ted with vgchangea s) of version 1.0 and 2.0

must have the consistency recovery set to NOMWC or NONE. Versions 1.0 and 2.0 do not support MWC forlonging to shared volume groups.

es belonging to version

the MWC is activatedv3, its MWC format is converted. Any subsequent activation of this volume group on previous

n of the entire logicald, activation of the

is happens during a Serviceguard rollingeleases might be enhanced to recognize the new MWC format; contact yourtative for up-to-the-moment information, or consult the Hewlett-Packard IT

Performance Improvements

The MWC allows a fast resynchronization of data following a system crash or failure.

The MWC keeps track of where I/O writes are occurring on the volume group and periodical

on the physical volume. Upon system reboot after crash, the operating system uses the Minconsistent data blocks quickly.

In HP-UX 11i v3, the MWC is larger in size than in previous releases. This leads to a better

To enable or disable MWC, use the -M y or -M n option with lvcreate or lvchange comma

more information on MWC, see lvcreate(1M) and lvch

Logical volumes belonging to shared volume groups (those activa

logical volumes be

With the September 2008 release of HP-UX 11i v3, LVM supports MWC for logical volum2.1 shared volume groups. This ensures faster recovery following a system crash.

Compatibility Note:

When a volume group containing a logical volume (created on pre-11i v3 releases) usingon HP-UX 11i

releases do not recognize the new MWC format and LVM performs a full resynchronizatiovolume. But if the logical volumes are marked NOMWC and the logical volume is synce

volume group on previous releases avoids full resynchronization. Note thupdate configuration. Previous rHewlett-Packard support represenResource Center site: http://itrc.hp.com(Americas and Asia Pacific) or http://europe.itrc.hp.com(Europe) for

orted on HP-UX 11i v2 with the patchPHKL_36244 or any superseding patch. With the patch PHKL_36244 or any superseding patch installed on HP-

O sizes (up to the extent size). As I/O requests do not need to be split into smallerchunks, the I/O throughput is improved by issuing reads and writes of larger sizes through LVM.

Note that the actual I/O size might be limited by the interface card.

For more information, seethe HP-UX System Administrator's Guide: Logical Volume Managementavailable athttp://docs.hp.com/en/oshpux11iv3.html#LVM%20Volume%20Manager

LVM patches.

NOTE: The new format of MWC introduced in HP-UX 11i v3 is supp

UX 11i v2, the new format of MWC is recognized and used to perform the recovery.

LVM Support for Large I/O Sizes

LVM now supports larger I/

Also see the The Next Generation Mass Storage Stackwhitepaper.

HP recommends you see the whitepaper for information on benefits and capabilities introduced in the New MassStorage Stack in HP-UX 11i v3.


18/35

18

Parallel Resynchronization of Stale Extents

command line options to the lvextend,

s the automatic

done by using the lvsynccommand. A new option,

onizes the specified logical volumes in parallel ratherthan serially. This reduces the time taken to synchronize the stale extents.

For more information, see lvextend(1M), lvmerge(1M) and lvsync(1M).

new command line option Tis introduced in the vgchangeVM activates a volume group. Normally, when a volume group is activated, all

p are attached sequentially. Activating the volume group using Toption

of the vgchangecommand attaches all the physical volumes in parallel. This reduces the time taken to activate

multiple volume groups with large number of physical volumes.

For more information, see vgchange(1M).

With September 2007 release of HP-UX 11i v3, LVM provides new

lvmerge,, and lvsynccommands that control the synchronization of stale extents.

A new option, s, is introduced in the lvextendand lvmergecommands, which suppresse

synchronization of stale extents. This synchronization can be

T, is introduced in the lvsynccommand, which resynchr

Parallel Open of Physical Volumes

In the September 2007 release of HP-UX 11i v3, acommand that controls the way L

the physical volumes of the volume grou


19/35

19

ts

Maximum Logical Volume Size Increased up to 16TB

LVM in HP-UX 11i v3 release supports logical volume of sizes up to 16 terabyte (TB) (versus 2TB in earlier

logical volume sizes must be a multiple of the extent size. Extent size is expressed in units of megabytes(MB) in the range of 1 to 256 and number of extents per logical volume is a decimal value in the range of 1 to

ts and 256MB extent size has the maximum supported size (16TB

The following are the advantages of 16TB logical volumes:

These logical volumes can be mirrored, striped, or striped and mirrored.

Existing logical volumes using 256MB extents that are 2TB or less can be extended up to 16TB.

Note: For the logical volumes to extend up to 16TB, physical volumes in volume group must have sufficient LVMolume Group Dynamic

Increased Limi

releases).

In LVM,

65535. Thus a logical volume with 65535 exten 256MB).

These logical volumes offer better scalability.

These logical volumes are supported in standalone, Shared LVM, and exclusive mode.

configuration space on the disk. To increase the LVM configuration space, see the LVM VLUN expansion (DLE)/vgmodifywhitepaper.

Compatibility Note

e. If a volume grouplarger than 2TB is created on HP-UX 11i v3, its activation and use is not recommended on any

previous HP-UX release. Though the volume group can be activated and used on earlier releases, the data beyond2TB is inaccessible.

2 with patchestches installed on aan be successfully

s

With the March and September 2008 releases of HP-UX 11i v3, LVM supports two new versions of volume group,Version 2.0 and Version 2.1 respectively. These new versions support configuration of larger number volumegroups, logical volumes, physical volumes, and several other volume group parameters. For more information onthe supported limits of Version 1.0 and Version 2.x volume groups, see the LVM Limits White Paper

Releases prior to HP-UX 11i v3 can only access data within the first 2TB of a logical volumwith logical volume

NOTE: Logical volume with size greater than 2TB and up to 16TB is supported on HP-UX 11i vPHCO_36744,PHKL_36745, and PHCO_36746. With these three patches or superseding paHP-UX 11i v2 system, a volume group with such a logical volume created on HP-UX 11i v3 cimported and activated.

Increased Limits for Version 2.x Volume Group

.

HP MirrorDisk/UX software is supported in a clustered environment with a maximum of two nodes configured forvolume group version 1.0 and 2.0. For volume group versions 2.1, HP MirrorDisk/UX software is supported in aclustered environment with a maximum of sixteen nodes configured.


20/35

20

Compact, Parsable Command Output

en enhanced to produce a compact and easily parsable output. A new option -F

is introduced in pvdisplay, vgdisplay ,andlvdisplay commands. The commands generate colon-separated fields as

ut can be split across multiple lines. The output

can include new keys or values in the future. The positioning of a given key and value pair is not

ed. If a key is deprecated, its associated value is set to NAM (key=NAM). HP recommends all user scripts

Examples

ogical volume name and field name as specified in the lvdisplay -Foutput, the following script

the value for the given key:

LVDI SPLAY_F_OUTPUT_KEY=$2;

LV_NAME | / usr/ bi n/ awk - F' : ' ' {f or ( i =1; i


21/35

21

`;

n");Pr ovi de a PV whi ch bel ong t o a VG\ n" ) ;

f or e r s) { my@pai r = ( ) ;

e ) ;

$pvdi spl ay_hash{$PVDI SPLAY_KEY}\ n";

t and checks if disk is under HP-UX LVM control

The ced to list additional

deta

The

-d

byte blocks from the

table physical volume, and displays thenumber of bad blocks that were relocated. These details are displayed in addition to other information.

to the end of the user data in 1024 byte

-l

Checks whether physical volume refers to a disk device under HP Logical Volume Manager (LVM) control.with many disks, use this

M disk, displays the start and end of the user data in 1

is supported starting with the March 2008 release of HP-UXember 2008 release of HP-UX 11i v3, this option is enhanced to

display whether the p otable or not.

Starting with the September 2007 release of HP-UX 11i v3, the pvdisplaycommand also displays the

proactive polling status of the physical volume. For more information, see the pvdisplay(1M) manual page.

Examples

# pvdisplay -d /dev/dsk/c22t6d2

- - - Physi cal vol umes - - -PV Name / dev/ dsk/ c22t6d2VG Name / dev/ vgnamePV St atus avai l abl e

my $pvdi spl ay_out = `pvdi spl ay - F $PV_NAME 2> / dev/ nul l

i f ( $? ! = 0) {pri nt("The PV $PV_NAME does not belong to any VG\pr i nt ( "exi t (1) ;

}

my @pvdi spl ay_pai r s = spl i t ( ' : ' , $pvdi spl ay_out ) ;

ach my $key_val ue ( @pvdi spl ay_pai

@pai r = spl i t ( ' =' , $key_val u$pvdi spl ay_hash{$pai r [ 0] } = $pai r [ 1] ;

}

pr i nt "

# get_pvinfo /dev/dsk/c11t0d5 vg_name

/ dev/ vgt est

pvdisplay Displays user data offse

pvdisplay command displays information about physical volumes in LVM. It is enhan

ils for a physical volume and to report if the disk is owned by LVM.

pvdisplay command has the following new options:

For each physical volume, displays the offset to the start of the user data in 1024

beginning of the physical volume, specifies if pv_path is a boo

Starting with the March 2008 release of HP-UX 11i v3, the offsetblocks is also displayed.

You can use this option on any disk in the system. For example, on a systemoption to determine which disks are marked for LVM usage and which are not.

-u

If physical volume is an LV

kilobyte blocks. This option11i v3. In the Sept

hysical volume is bo


22/35

22

25

e ( Mbyt es) 41279

n1024

ume is owned by LVM, do the following:

/c4t3d0

VM, do the following:

Bootable=yes

Group Scanning and Supports Persistent DSF

introduced in March

200 ds entries for volume

grou vers volume group

info searching for LVM disks.

-f

With this option vgscanworks on per volume group basis rather than scanning for

This option replaces any existing entries related to these volume gr

with updated entries. If the specified volume groups are missing, vgscanadds the

/etc/lvmtab_p.

Note: vgscan -fdoes n

command and does not report unconfigured volume groups.


23/35

s,

23

With the above option vgscanprovides the following improvements:

corrected without

in a cumbersome way

.

d /etc/lvmtab_p, such as

the -koption, vgscancan recover activated volume groups in /etc/lvmtab and /etc/lvmtab_p quickly (in

ds), even if the system has a large number of LVM disks configured.

lume DSFs in /etc/lvmtab with therdware probe of all devices.

ormat i on

Vol ume Gr oupVol ume Gr oup

l ume Gr oupVol ume Gr oup

i s not par t of a Vol ume Gr oupVol ume Gr oup

ol ume Gr oup"/ dev/ dsk/ c20t 13d3" i s not part of a Vol ume Gr oup

i on

t o resync t he i nf or mat i on on t he di sk. ** *

ev/vg01, overwrite the existing physical volume DSFs in /etc/lvmtab with thet be activated, or this

/etc/lvmtab with the

re probing of all devices.

/ c7t 0d1" : Thel ume at t ached t o t hi s

vgscan: Warni ng: coul dn' t quer y physi cal vol ume " / dev/ dsk/ c9t 0d1" : Thespeci f i ed path does not cor r espond t o physi cal vol ume at t ached t o t hi svol ume groupvgscan: Warni ng: coul dn' t quer y al l of t he physi cal vol umes.vgscan: The physi cal vol ume "/ dev/ di sk/ di sk25_p2" i s al r eady recor ded i n t he"/ et c/ l vmt ab" f i l e.vgscan: The physi cal vol ume " / dev/ di sk/ di sk30" i s al r eady r ecorded i n t he"/ et c/ l vmt ab" f i l e.vgscan: The physi cal vol ume " / dev/ di sk/ di sk31" i s al r eady r ecorded i n t he"/ et c/ l vmt ab" f i l e.Physi cal Vol ume "/ dev/ di sk/ di sk26" cont ai ns no LVM i nf ormati on

With the -foption, an incorrect volume group entry in /etc/lvmtab or /etc/lvmtab_p can be

modifying the rest of /etc/lvmtab or /etc/lvmtab_p. Previously, this could only be done

using vgexportand vgimportto recover lvmtab or lvmtab_p entries for the root volume group With the -foption, a user can specify the ordering of volume groups in /etc/lvmtab an

placing the boot volume group first.

With

secon

Examples

For the volume group /dev/vg01, overwrite the existing physical vophysical volume DSFs found belonging to /dev/vg01 during a ha

# vgscan -f /dev/vg01

Physi cal Vol ume "/ dev/ dsk/ c2t 0d0" cont ai ns no LVM i nfCoul dn' t st at physi cal vol ume "/ dev/ dsk/ c0t 0d0":I nval i d argumentPhysi cal Vol ume "/ dev/ dsk/ c13t 0d5" i s not part of aPhysi cal Vol ume "/ dev/ dsk/ c11t 0d5" i s not part of a

Physi cal Vol ume "/ dev/ dsk/ c20t 0d6" i s not par t of a VoPhysi cal Vol ume "/ dev/ dsk/ c17t 0d6" i s not part of aPhysi cal Vol ume " / dev/ dsk/ c17t 1d0"Physi cal Vol ume "/ dev/ dsk/ c20t 1d0" i s not part of aPhysi cal Vol ume "/ dev/ dsk/ c17t 13d3" i s not par t of a VPhysi cal Vol umePhysi cal Vol ume "/ dev/ dsk/ c17t 13d6" cont ai ns no LVM i nf ormat i onPhysi cal Vol ume "/ dev/ dsk/ c20t 13d6" cont ai ns no LVM i nf ormat

** * LVMTAB has been updat ed successf ul l y.*** Do the f ol l owi ng#1. vgchange - a y*** #2. l vl nboot - R

For the volume group /d

physical volume DSFs used in kernel memory. The volume group /dev/vg01 mus

command will fail.

# vgscan -k -f /dev/vg01

For the volume group /dev/vg01, overwrite the existing physical volume DSFs in

physical volumes persistent DSFs found belonging to /dev/vg01 during a hardwa

# vgscan -N -f /dev/vg01

vgscan: Warni ng: coul dn' t quer y physi cal vol ume " / dev/ dskspeci f i ed path does not cor r espond t o physi cal vo

vol ume group


24/35

physi cal vol ume "/ dev/ di s

24

k/ di sk27":

r t of a Vol ume Group/ di sk80":

Vol ume Gr oupof a Vol ume Gr oup

f a Vol ume Gr oup"/ dev/ di sk/ di sk96" cont ai ns no LVM i nf or mat i oneen updated successf ul l y.

b_p files for volume groups activated since the last boot. For activatedh persistent and legacy DSFs. Report all physical volume persistent andvolume groups.

# mv /etc/lvmtab_p /etc/lvmtab_p.BCK

t par t of a Vol ume Gr oup" / dev/ dsk/ c5t 12d0" i s not par t of a Vol ume Gr oup

e Gr oup.l ume Gr oup.

e pr ocess.k35_p2d0s2

Vol ume Gr oup.ch t hese Physi cal Vol umes t o a Vol ume Gr oup.

por t command to compl et e the process.sk38_p2

/ dev/ dsk/ c4t 0d0s2

Vol ume Gr oup.ch t hese Physi cal Vol umes t o a Vol ume Gr oup.or t command t o compl et e the process.

Fol l owi ng Physi cal Vol umes bel ong t o one Vol ume Gr oup.Unabl e to mat ch t hese Physi cal Vol umes t o a Vol ume Gr oup.Use t he vgi mpor t command t o compl et e t he pr ocess./ dev/ di sk/ di sk45/ dev/ dsk/ c4t 14d0/ dev/ dsk/ c5t 14d0

Fol l owi ng Physi cal Vol umes bel ong t o one Vol ume Gr oup.Unabl e to mat ch t hese Physi cal Vol umes t o a Vol ume Gr oup.Use t he vgi mpor t command t o compl et e t he pr ocess.

Coul dn' t st atI nval i d argument

me "/ dev/ di sk/ di sk36" i s not paPhysi cal Vol uCoul dn' t st at physi cal vol ume "/ dev/ di sk

entI nval i d ar gumCoul dn' t st at physi cal vol ume "/ dev/ di sk/ di sk81":I nval i d argumentPhysi cal Vol ume "/ dev/ di sk/ di sk82" i s not par t of a

3" i s not par tPhysi cal Vol ume "/ dev/ di sk/ di sk8

Physi cal Vol ume "/ dev/ di sk/ di sk93" i s not par t oPhysi cal Vol ume*** LVMTAB has b*** Do the f ol l owi ng t o resync t he i nf or mat i on on t he di sk. ** *#1. vgchange - a y*** #2. l vl nboot R

Recreate the /etc/lvmtab and /etc/lvmtaand deactivated volume groups, use bot

s belonging to unconfiguredlegacy DSF

# mv /etc/lvmtab /etc/lvmtab.BCK

# vgscan B

Cr eat i ng "/ et c/ l vmt ab"."/ dev/ dsk/ c4t 12d0" i s noPhysi cal Vol ume

mePhysi cal Vol u

Fol l owi ng Physi cal Vol umes bel ong t o one Vol umt o a VoUnabl e t o mat ch t hese Physi cal Vol umes

t command to compl et e t hUse t he vgi mpor/ dev/ di sk/ di s/ dev/ dsk/ c4t 1/ dev/ dsk/ c5t 1d0s2

Fol l owi ng Physi cal Vol umes bel ong t o oneUnabl e t o matUse t he vgi m/ dev/ di sk/ di

/ dev/ dsk/ c5t 0d0s2

ysi cal Vol umes bel ong to oneFol l owi ng PhUnabl e t o matUse t he vgi mp/ dev/ di sk/ di sk48/ dev/ dsk/ c4t 5d0/ dev/ dsk/ c5t 5d0


25/35

25

k46k473d03d00d0

d wi t h any Physi cal Vol umes.

e conf i gur ed i n t he syst em.i on on di sk.

/ di sk44":

k/ di sk49":

sk/ di sk40":

physi cal vol ume " / dev/ dsk/ c5t 4d0" :

" :

physi cal vol ume " / dev/ dsk/ c0t 0d0" :ment

e Gr oup.o a Vol ume Gr oup.

or t command t o compl et e the process.k33

1d0

ysi cal Vol umes bel ong t o one Vol ume Gr oup.

/ dev/ dsk/ c4t 9d0

Vol umes.ol umes.

al Vol umes.ol umes.

nual page.

vgcfgrestore Enhanced Listing of Backup File Contents

The vgcfgrestorecommand restores the LVM configuration data from a default (-noption) or alternate (-foption)

configuration backup file to a physical volume, or it displays the configuration backup file content ( -loption).

In HP-UX 11i v3, the vgcfgrestore-lcommand provides the following additional information from the configuration

backup file when invoked together with the -voption:

For each physical volume, display the disk size in kilobytes

/ dev/ di sk/ di s/ dev/ di sk/ di s/ dev/ dsk/ c4t 1/ dev/ dsk/ c5t 1/ dev/ dsk/ c4t 1/ dev/ dsk/ c5t 10d0

The Vol ume Gr oup / dev/ vg01 was not mat chen cr eat ed successf ul l y.*** LVMTAB has bee

ar*** I f PV l i nks*** Do the f ol l owi ng t o r esync i nf ormatchange - a y** * #1. vg

*** #2. l vl nboot - Rt c/ l vmt ab_p".Cr eat i ng "/ e

Coul dn' t st at physi cal vol ume " / dev/ di skI nval i d argument

t physi cal vol ume " / dev/ di sCoul dn' t st aI nval i d argument

physi cal vol ume "/ dev/ diCoul dn' t st atI nval i d argumentCoul dn' t st atI nval i d argumentCoul dn' t st at physi cal vol ume "/ dev/ dsk/ c4t 4d0

I nval i d argumentCoul dn' t stI nval i d ar gu

at

Fol l owi ng Physi cal Vol umes bel ong t o one Vol umUnabl e t o mat ch t hese Physi cal Vol umes tUse t he vgi mp

i s/ dev/ di sk/ d/ dev/ dsk/ c2tFol l owi ng PhUnabl e t o mat ch t hese Physi cal Vol umes t o a Vol ume Gr oup.Use t he vgi mpor t command t o compl et e t he pr ocess.

/ dev/ di sk/ di sk37/ dev/ di sk/ di sk36/ dev/ dsk/ c4t 2d0/ dev/ dsk/ c5t 2d0

/ dev/ dsk/ c5t 9d0

The Vol ume Gr oup / dev/ vgl 21 was not mat ched wi t h any Physi calThe Vol ume Gr oup / dev/ vgl 1 was not mat ched wi t h any Physi cal VThe Vol ume Gr oup / dev/ vgL202 was not mat ched wi t h any Physi cThe Vol ume Gr oup / dev/ vg21 was not mat ched wi t h any Physi cal V** * LVMPTAB has been cr eat ed successf ul l y.

For more information, see the vgscan(1M) ma


26/35

26

he lvdisplay(1M) manual page for more information). For each volume group, display max_pv, max_pe, and max_lv.

e group version also. For a

x volume group, it also displays the maximum size to which the volume group can grow.

V 0. conf "

PV Type Size (kb) Start (kb) PVkeye 34643968 2912 0

ax_lv 255 vg_version 1.0

onf / vg02. conf "

- - - -PV Type Si ze ( kb) St art ( kb) PVkey

1 vg_version 2.0 vg_size 20t

error descriptions,

X 11i v3, LVM supports EMT and

Long Host Name Support for Display Commands

Shared LVM (SLVM) allows a volume group to be shared across different nodes within a Serviceguard cluster. In

cipating in the HP Serviceguard

presented as Server and all other nodes as Client.

In HP-UX 11i v3 the vgdisplay command has been enhanced to support large host names of more than 32

Note that though LVM supports host name length up to 255 characters, the host name displayed depends on thehost name length supported by HP Serviceguard.

For more information, see the Node and Host Name Sizes on HP-UX

For each physical volume, display the starting block number (KB) of the user data

For each physical volume, display the PVkey (See t

With the March 2008 release of HP-UX 11i v3, the voption display the volum

Version 2.

Example

# vg cfgrestore l -v -n vg00

i gur atol ume Group Conf i on i nf ormat i on i n " / et c/ l vmconf / vg0VG Name / dev/ vg00- - - - Physi cal vol umes : 1 - - - -

di sk95_p2 Bootabl

max_pv 16 max_pe 4238 m

For a Version 2.0 volume group:

# vgcfgrestore -l -v -n vg02Vol ume Gr oup Conf i gur at i on i nf ormat i on i n " / et c/ l vmcVG Name / dev/ vg02- - - - Physi cal vol umes : 1

di sk50 Non- Boot 71687369 1024 0

max_ pv 511 max_ pe 2097152 max_ l v 51

Error Management Technology (EMT)

Error Management Technology (EMT) provides an online searchable repository containing

probable causes and recommended actions on systems running HP-UX. With HP-ULVM error messages are available in the EMT repository.

this environment, the vgdisplay command displays the host names of the nodes parti

cluster. One of the host names is re

characters, (as in prior releases) up to 255 characters.

whitepaper. The whitepaper providesinformation on supported host name lengths.

Example

Display the hostname information (machine-server and machine-client)


27/35

# vgdi spl ay - v / dev/ vgt est 1

27

e Accessat us avai l abl e, shar ed, ser ver

255V 1

116

1per PV 8750

2( Mbyt es) 8

874950

Fr ee PE 8699TTT e 0

ver Server

t Client

ogi cal vol umes -

/ dev/ vgt est 1/ l vol 1l abl e/ syncd

( Mbyt es)t LE 50

l vol umes - - -PV Name / dev/ dsk/ c4t13d0

Fr ee PE 8693

Command

With the March 2008 release of HP-UX 11i v3, LVM provides a new command, lvmadm, which displays the

p limits for specified volume group versions supported by LVM.

HP-UX 11i v3, this command is enhanced to support a new option, -l, which

etc/lvmtab and /etc/lvmtab_p files for the specified volume group version. This

ning the command st r i ngs / et c/ l vmt abor st r i ngs / et c/ l vmt ab_p.

Examples

To view the supported limits for all volume group versions supported by LVM:

# l vmadm - t- - - LVM Li mi t s - - -VG Versi on 1. 0Max VG Si ze ( Tbyt es) 510Max LV Si ze ( Tbyt es) 16Max PV Si ze ( Tbyt es) 2

- - - Vol ume gr oups - - -VG Name / dev/ vgt est 1

r i t r ead/ wr i t eVG WVG StMax LVCur LOpen LVMax PV

Cur PV 1Act PVMax PEVGDAPE Si ze

Tot al PEAl l oc PE

ot al PVG 0ot al Spar e PVs 0otal Spare PVs i n us

machine1-ser

lienmachine2-c

- - - L - -LV Name

atusLV St avai400LV Si ze

Cur r enAl l ocat ed PE 50Used PV 1

- - - Phys i ca

PV St atus avai l abl e

Tot al PE 8749

Aut oswi t ch On

lvmadm

volume grou

In the September 2008 release of

/displays the configuration in the

serves as a replacement for run


28/35

28

2562552552255327686553565535256

2. 02048256165121111

2621443355443216777216256

2. 120482566048

2047

Max St r i pes 5112621443355443216777216

s) 256

To display the Version 1.0 volume groups and the corresponding physical volumes as contained in the LVM

# l vmadm - l 1. 0

v/ vg00_p2

For more information, see the lvmadm(1M) manual page.

vgremove Removal of Volume Group Device Files

With the March 2008 release of HP-UX 11i v3, LVM supports a new -Xoption, which removes the volume group

device files in /dev.

For more information, see vgemove(1M).

vgdisplay Displays More Information

Max VGsMax LVsMax PVsMax Mi r r orsMax St r i pesMax St r i pe Si ze ( Kbyt es)Max LXs per LVMax PXs per PVMax Ext ent Si ze ( Mbyt es)

VG Ver si onMax VG Si ze ( Tbyt es)Max LV Si ze ( Tbyt es)Max PV Si ze ( Tbyt es)Max VGsMax LVs 5Max PVs 5

5Max Mi r r orsMax St r i pes 511Max St r i pe Si ze ( Kbyt es)Max LXs per LVMax PXs per PVMax Ext ent Si ze ( Mbyt es)

VG Ver si onMax VG Si ze ( Tbyt es)Max LV Si ze ( Tbyt es)Max PV Si ze ( Tbyt es) 1Max VGs 2Max LVsMax PVs 2048Max Mi r r or s 5

Max St r i pe Si ze ( Kbyt es)Max LXs per LVMax PXs per PV

Max Ext ent Si ze ( Mbyt e

configuration file:

- - - Ver si on 1. 0 vol ume gr oups - - -VG Name / dePV Name / dev/ di sk/ di sk34


29/35

With the March 2008 release of HP-UX 11i v3, vgdisplaydisplays the volume g

29

roup version, the maximum size to

which the volume group can grow and the maximum number of physical extents.

LVM supports a new Voption, which displays the volume

formation for all volume groups of the specified version.

/ dev/ vg00r ead/ wr i t eavai l abl e255441611423828

42280292199

i n use 01.0

464m

08

groups on the system:

/ dev/ vg00r ead/ wr i t eavai l abl e25544161

4238

8

Fr ee PE 2199Tot al PVG 0Tot al Spar e PVs 0Tot al Spar e PVs i n use 0VG Versi on 1. 0VG Max Si ze 542464mVG Max Ext ent s 67808

VG Name / dev/ vg01VG Wr i t e Access r ead/ wr i t eVG St atus avai l abl e

With the September 2008 release of HP-UX 11 iv3,

group in

Examples

# vgdi spl ay vg00

- - - Vol ume gr oups - - -VG NameVG Wr i t e AccessVG St at usMax LVCur LVOpen LVMax PVCur PVAct PVMax PE per PVVGDAPE Si ze (Mbyt es)

Tot al PE 2Al l oc PEFr ee PE

Tot al PVG 0Tot al Spar e PVs 0Tot al Spar e PVsVG VersionVG Max Size 542

8VG Max Extents 67

eTo display all Version 1.0 volum

# vgdi spl ay - V 1. 0

- - - Vol ume gr oups - - -VG NameVG Wr i t e AccessVG St at usMax LVCur LVOpen LVMax PVCur PVAct PV 1Max PE per PVVGDA 2PE Si ze ( Mbyt es) 8

2Tot al PE 42

Al l oc PE 2029


30/35

30

2551116111750124

175011017491

Tot al Spar e PVs i n use 0

VG Max Si ze 1120064m 280016

For more information, see the vgdisplay(1M) manual page.

y created the directory for the volume group

group using the mkdi r and mknodcommands.

upport automatic

name/ gr oup

file. If the file is not present, the command creates one.

Using the automatic creation of volume group control file feature, it is possible to have a shared volumedifferent nodes of a cluster. When creating shared volume groups, HP

with a minor number that is free on all nodes of the cluster.

-UX 11i v3, the pvmovecommand is enhanced to have several new

ysical extents.physical volume.

ume.

d mirrored logical volumes. Support for a new option, -p, to preview physical extent movement details without performing the move.

Examples

To preview the movement of first physical extent from source physical volume /dev/dsk/c5t7d0 to any

physical volume in the volume group:

# pvmove -p /dev/dsk/c5t7d0:0

- - - Pr evi ew of t he Rel ocat i on Oper at i on - - -

Sour ce PV Sour ce PE Dest i nat i on PV Dest i nat i on PE

Max LVCur LVOpen LVMax PVCur PVAct PVMax PE per PVVGDAPE Si ze (Mbyt es)

Tot al PEAl l oc PEFr ee PE

Tot al PVG 00Tot al Spar e PVs

VG Versi on 1. 0

VG Max Ext ent s

Auto Creation of Volume Group File

Prior to the March 2008 release of HP-UX 11i v3, the user manuall

and the device file named

With the March 2008 release of HP-UX 11i v3, the vgcr eat eand vgi mport commands screation of the directory of the volume group and the volume group control file / dev/ vgregardless of the volume group version. The commands retain support for manually created

/ dev/ vgname/ groupFor more information, see the vgcreate(1M) and vgimport(1M) manual pages.

NOTE:group with different minor numbers onrecommends that you manually create the group file

pvmove Enhancements

In the September 2008 release of HP

features:

Support for moving a range of ph Support for moving extents from the end of aSupport for moving extents to a specific location on the destination physical vol

Support for moving the physical extents from striped logical volumes and stripe


31/35

/

31

5 to 100 from source physical volume /dev/disk/disk10 to the desired

22

/dev/disk/disk10 to another physical volume:

stination physical volume /dev/disk/disk22, starting at

physical extent 50 on /dev/disk/disk22:

# pvmove /dev/disk/disk10 /dev/disk/disk22:50

dev/ dsk/ c5t7d0 0 / dev/ dsk/ c5t 7d0 4

To move extents in the range 2

destination physical volume /dev/disk/disk22:

# pvmove /dev/disk/disk10:25-100 /dev/disk/disk

To move the last 5 extents from

# pvmove e 5 /dev/disk/disk10

To move extents to a specific location on the de


32/35

32

Miscellaneous

s

ry 2007 release of HP-UX 11i v3, LVM commands were enhanced to avoid overwriting LVMconfiguratio n p sical volumes that already belong to a volume group and are recorded in the

s not present in

/etc/lvmt eded even though the corresponding physical volume belonged to a volume group recorded in

ary 2007 release of HP-UX 11i v3 so that pvcreate

or is the same for both legacy and persistent DSFs of the physical volume.

In addition to pvcreate the following commands have been enhanced similarly:

t1d0 and

c9t1d0.

ases prior to HP-UX 11i v3. But on HP-UX 11i v3

pvcreate fails and prints the following:

"/ dev/ dsk/ c9t 1d0" i s al r eady pr esent i n t he"/ et c/ l vmt ab" f i l e and r epr esent s an al t er nat e l i nk to "/ dev/ dsk/ c6t 1d0".

Mirror Disk Installation No Longer Requires a Reboot

With the HP-UX 11i v3 release, the Mirror Disk product installation is much easier. Unlike previous releases, Mirror Diskproduct installation does not require a reboot.

Commands enhanced to prevent misconfigurations through alternate link

Starting with the Februan data o hy

/etc/lvmtab or /etc/lvmtab_p.

In prior releases, pvcreatewith -f (force) option on an alternate physical volume path that wa

ab succe

/etc/lvmtab. This behavior has been changed with the Febru

fails in such situations. The behavi

vgcreate, pvremove, vgimport, and vgchgid.

Example

As an example, consider a multipathed physical volume whose DSFs are /dev/dsk/c9

/dev/dsk/c6t1d0. Let the volume group be configured with only /dev/dsk/

pvcreatef /dev/rdsk/c6t1d0 succeeds on rele

pvcreat e: The physi cal vol ume


33/35

33

Agi

hysical location of theLUN remains the

ved from one HBA to another, moved from one switch/hub port to another,presented via a different target port to the host, or configured with multiple hardware paths. Also referred

as persistent LUN binding.

Agile View

The representation of LUNs using lunpath hardware paths, LUN hardware paths, and persistent DSFs,

Hardware Path

ies of numbers representing the physical or virtualized location of a device. The path is a sequenceight not correspond to

ght represent only a handle to a device rather than a physical path

Leges prior to HP-UX 11i v3.

xists in releases prior to HP-UX 11i v3. It is composed of aseries of bus-nexus addresses separated by / leading to the HBA; beneath the HBA, additional addresselements (such as domain, area, port, target, and LUN) are separated by .. The string

y DSFded in the files

minor name and file name, such as / dev/ dsk/ c2t 3d4.

LU

refers to an end storage device such as a disk, tape, floppy, or CD. This is thegical unit itself and does not represent the path to the logical unit.

The physical hardware path leading to a SCSI logical unit. A SCSI LUN can have more than one lunpath.

Persistent DSF

A DSF conforming to the naming model introduced in HP-UX 11i v3 to support agile addressing. The

device file name contains an instance number, such as / dev/ di sk/ di sk#, and the minor number has

no hardware path information.

MetadataThe on-disk structures that LVM uses to manage a volume group. This space is not available forapplication data.

Glossary

le Addressing

The ability to address a LUN with the same device special file regardless of the pLUN or the number of paths leading to it. In other words, the device special file for asame even if the LUN is mo

to

introduced in HP-UX 11i v3.

DSFDevice Special File. A file associated with an I/O device. DSFs are read and written the same asordinary files, but requests to read or write are sent to the associated device.

A serof I/O addresses that share a hierarchical relationship. The address elements m

physical hardware addresses, and mito it.

acy ViewThe representation of legacy hardware paths and legacy DSFs, as in releas

Legacy Hardware PathThe representation of a hardware path as it e

0/2/1/0.1.4.0.0.2.7 is an example of a legacy hardware path.

LegacA DSF with the hardware path information such as SCSI bus, target, and LUN embed

N

A SCSI logical unit. Thislo

Lunpath


34/35

34

h type of volume groupassigned a version to distinguish from another type, based on the metadata format maintained by

LVM.

Volume group versionBased on the metadata format, LVM supports three types of volume groups. Eac

is


35/35

For More Information

g documents on HP documentation website:To learn more about some of the LVM features, see the followin

http://docs.hp.com(Use search with the given name of the whitepaper)20Volume%20Managerhttp://www.docs.hp.com/en/oshpux11iv3#LVM%

SNOR)

nsion (DLE)/vgmodify LVM Volume Group Quiesce/Resume

ume group configuration from legacy to agileon HP documentation website:

SLVM Single-Node Online Reconfiguration (SLVM

tLVM Online Disk ReplacemenDynamic LUN expa

LVM Volume Group

LVM Version 2.0 Volume Groups in HP-UX 11i v3

LVM Limits White Paper

To learn more about configuring LVM and migration of LVM volnaming model, see the following documents

http://docs.hp.com(Use search with the given name of the whitepaper)http://www.docs.hp.com/en/oshpux11iv3#LVM%20Volume%20Manager

ide Logical Volume Management

document on HP20Area%20Management

LVM Migration from Legacy to Agile Naming Model

HP-UX System Administrator's Gu

To learn more about the agile view and the new mass storage stack, see the followingdocumentation website: http://docs.hp.com/en/netsys.html#Storage%

rage Stack

t supported node and host name sizes on HP-UX, see the following document on HP

The Next Generation Mass Sto

To learn more aboudocumentation website: http://www.docs.hp.com/en/oshpux11iv3#White%20Papers

Node and Host Name Sizes on HP-UX

Call to Action

HP welcomes your input. Please give us comments about this whitepaper, or suggestions through our technicaldocumentation feedback website: http://docs.hp.com/en/feedback.html.

2008 Hewlett-Packard Development Company, L.P. The informationcontained herein is subject to change without notice. The only warranties forHP products and services are set forth in the express warranty statementsaccompanying such products and services. Nothing herein should be construedas constituting an additional warranty. HP shall not be liable for technical oreditorial errors or omissions contained herein.

35
http://www.docs.hp.com/en/oshpux11iv3%23White%20Papershttp://docs.hp.com/en/feedback.html.http://docs.hp.com/en/feedback.html.http://www.docs.hp.com/en/oshpux11iv3%23White%20Papers

hpux lvm new features

Documents

volume group scanning

volume group settings

group quiesce

hpux lvm new features

v3lvm new features

device online replacement

hpux lvm control

slv online reconfiguration