Veritas Volume Manager

VM daemons

vxconfigd

Maintains system configuration in the kernel & disk (private region). If the daemon is stopped it does not disable any configuration state loaded into the kernel, it only affects the ability to make configuration changes until vxconfigd is restarted.

It can be in three states:
Enable = Normal mode
Disable = Most operations cannot be be used
Booted = Normal startup while using boot disk group

vxrelocd Monitors for failure events and relocates failed subdisks
vxconfigbackupd Used to backup configuration chnages, the files created can be used with vxmake to restored lost groups.
vxnotify Display veritas volume manager events used with the vxconfigd daemon
vxcached used to administer a cache object that is assoicated with volumes that have one or more space-optimized snapshots. When the usage of a cached volume reaches the high watermark vxcached autom,atically grows the cache volume if required and configured
vxattachd Monitors VxVM for disks being attached and reattaches a detached site if the disks belongs to that sire become accessible, the daemon monitors vxnotify command and waits for a failed disk, when the disk is attached vxattachd attempts to online the disk, if successful it then starts recovery using vxrecover
vxdbd Handles comunication to and from the storage foundation product, it uses port 3233
Kernel Info
Kernel States

The kernel can be in three states:

Enabled - both private and public regions are accessible
Disabled - no private or public regions are accessible
Detached - only private regions are accessible


VM utilities

 VxVM debug

vxconfigd -k -m enable -x <debug option>

-x log                                log to /var/vxvm/vxconfigd.log
-x logfile = <name>              log to filename
-x syslog                            log to syslog
-x timestamp                      date and timestamp every entry
-x tracefile=name                log all possible tracing to file

vxiod

The vxiod utility starts, stops, or reports on VERITAS Volume Manager (VxVM) I/O daemons. An I/O daemon provides a process context for performing I/O in VxVM.Manage extended disk i/o & handles dirty regions, logging

vxiod set <number> = set number of runnning viod daemon

Note: when run on its own it displays # of vxiod daemons that are running.

vxdctl [option]

The vxdctl utility manages aspects of the state of the volume configuration daemon vxconfigd and also manages aspects of configuration for bootstrapping the rootdg disk group.

mode = what mode the vxconfigd is running in
enable = enable the vxconfigd daemon (reread the db)
disable = disable the vxconfigd daemon
stop = kill the vxconfigd daemon  (Use 'vxconfigd -k -m enable' to start again)
license [init] = print out license info or reread licenses
support = display version and components
list = display entries in /etc/vx/volboot
init [dmp] = recreate /etc/vx/volboot

Note: when in disabled or stopped mode no VX commands will be able to run

vxinstall

Install volume manger (use /etc/vx/disk to exclude any disks or controllers)

vxdbdctl used to stop/start the vxdbd daemon, you can also use status to obtain the status of the daemon

Disk Regions

Disk Regions
Private

This is were veritas holds the meta data regarding the disk. A copy of the configuration database is copied to each private region within the disk group. Veritas will try and keep 5 copies of the configuration database.

See below for where the slice is kept

Once the private region is created it cannot be changed, thus if it fills up then you have to reinitize the disk with a new private region length, see disks -> initialize disk

Public

This is the area that will store the users data.

See below for where the slice is kept

Disk Layouts
Sliced Disk layout

private and public region slices are on separate partitions, this type of disk is not suitable for moving between different O/S's but are suitable for boot partitions

Can be converted to CDS

format=sliced
CDS (Cross-platform Data Sharing)

private and public regions are one slice , this type is suitable for moving between different O/S's but not suitable for boot partitions.

format=cdsdisk
Simple

Private and public are the same partition but continuous

Can be converted to CDS

format=simple
None

No partitioning

n/a

VxVM Configuration Database

Public/Private partitions vxdisk list <disk> |egrep -i '^public|^private'
DB size # the size of the configuration db
vxdg list <group> | grep permlen
DB location vxdisk list <disk> | grep -i config - db location

File Locations

vxinstall has not be run /etc/vx/reconfig.d/state.d/install-db
Host ID's /etc/vx/volboot
backup config files (vxconfigbackupd) /etc/vx/cbr/bk

Note: you must have at least run vxconfigbackup once, otherwise the directory does not exist or you have specified another location.
delete or deported disk group config files /etc/vx/dgcfg/deport
All commands logs /var/adm/vx/cmdlog
/var/adm/vx/veacmdlog
Licenses /etc/vx/licenses/lic
Imported disk groups info

/var/vxvm/tempdb

Note: to clear the tempdb file:
vxconfigd -k -x cleartempdir - clear the /var/vxvm/tempdb

vxconfigd log file /var/adm/vx/vxconfigd.log

Backup & Restore

Backup

/usr/lib/vxvm/bin/vxconfigbackup -l /var/vxvm/backups

-l = location where to store backup

Restore (precommit/commit)

vxconfigrestore -p <group>

# either one of the below after the precommit
vxconfigrestore -d <group>
vxconfigrestore -c <group>

Note:
-p = when you want to check that the restore is correct (use vxprint to check)
-d = abort the precommit
-c = commit the precommit

Disks

Initialize disk

## increase the private region size default 1024
vxdisksetup -i c2t0d0 privlen= 2048

## change the default of a disk
vxdisksetup -i <device> format=sliced   - initialized a disk as a sliced disk

Note: format can be either sliced, simple, cdsdisk or none (see above - Disk Regions)

Uninitialize disk vxdiskunsetup -C c2t0d0
Disk Information vxdisk -g <group> list <disk>
vxdisk -s list
Resize a LUN vxdisk -g <group> resize <disk> length=8G
Add a disk slice to volboot vxdctl add disk <device> type=simple
Add a disk slice vxdisk –f <device> type=simple
Add a disk vxdiskadd c1t0d0 or c1 (all disk on controller)
vxdisksetup -i <device>
Remove a disk totally from VM vxdisk rm <device>
Remove a disk from a volume vxdg -g <group> rmdisk <diskname>
Remove a disk slice from VM vxdctl rm disk <device>
Clear any host ID flags vxdisk clearimport <disk name>
Renaming a disk vxedit -g <disk> rename <old disk name> <new disk name>
Move disk to different disk group vxdg move <source dg> <target dg> <disk>

Note: you must always have one disk in a disk group, you cannot move a disk that is used by a volume
Offline a disk

vxdisk offline <dev name>

Note: disk must not be in a disk group

Online a disk

vxdisk online <dev name>

Note: disk must have a private region otherwise you need to initialise the disk

Hot spare vxedit –g <group> set spare=on <disk>
NoHotUse vxedit –g <group> set nohotuse=on <disk>
Turn off failing flag vxedit -g <group> set failing=off <disk>
Encapsulate a disk vxdisk define c0t0d0s0 type=nopriv
Reattach disk (SAN)

vxreattach [-br | -c <device>]

-b = Background process
-r = Recover volumes
-c = Checks to see if reattach is possible

Discover new disks vxdisk scandisks [new | fabric ]
Disk Comment vxedit -g <group> set comment="......." <disk>
Private region problem

## Here i am fixing a "online altused" issue but search the web for different issues as there are a number of ways to the same thing, just remember the vxprivutil command

## obtain the disks private region tag 15
prtvtoc /dev/rdsk/c2t0d0s2

## dd the content sto a file
dd if=/dev/rdsk/c2t0d0s3 of=/tmp/c2t0d0s3_privreg

## run the below script
fix_script

## if now errors reported above and in the file, then dd back to disk
dd if=/tmp/c2t0d0s3_privreg.good of=/dev/rdsk/c2t0d0s3

## repeat for any other bad disks


Disk Group

Create a disk group vxdg init group <disk>=<device>
vxdg init <group> <disk>=<device> cds=off    - initialize a non-cds disk group
Remove a group vxdg destroy <group>
Add a disk to a group vxdg –g <group> adddisk <disk>=<device>
Remove a disk from a group vxdg –g <group> rmdisk <disk>
Replace failed disk

vxdg -k -g <group> adddisk <disk>=<device>

-k = forces vxvm to take media name of the failed disk & assign it to the new disk

Import a group

vxdg import <group>
vxdg -n <new-group-name> import <old-group-name>
vxdg -C import <group>

Import group (clear any flags)

vxdg import -C <group>

-C - clears any exist host flags

Deport a group vxdg deport <group>
vxdg -n <new-group-name> deport <old_group_name>
List no hot use on disk vxdg –g <group> nohotuse
List spare space on disk vxdg –g <group> spare
Display free space vxdg –g <group> free
Backup disk group (vxvm 4.0) vxconfigbackup
Restore disk group (vxvm 4.0) vxconfigrestore [-p|-d|-c]
Diskgroup Version vxdg list <group> | grep -i version
Upgrade disk version vxdg upgrade <group>                       - upgrade to current version
vxdg -T 50 upgrade <group>                 - upgrade to version 50
vxdg -T 50 init <group> <disk>=<device>    - creater new group @ version 50
Boot/Default DG vxdg bootdg
vxdg defaultdg

vxdctl defaultdg <group> - set defaultdg

Volume

Adding mirror to root /etc/vx/bin/vxrootmir <alternate> create rootvol, swap vol,
Create a simple volume

vxassist make <volume> <size> <disk>
vxassist -g <group> make <vol> <size> !ctrl:c2     - don't use controller 2

vxassist make <volume> <size> <disk> layout=[stripe-mirror|concat-mirror|mirror-concat|mirror-stripe]

stripe-mirror = layered volume
concat-mirror = layered volume
mirror-concat = non-layered volume
mirror-stripe = non-layered volume

Mirror a simple volume vxassist mirror <volume> <disk>
vxassist -g <group> remove mirror <vol> !disk01    - remove the disk01 mirror
Create a stripped volume

vxassist make <volume> <size> layout=stripe

vxassist -g <group> -o ordered make <vol> <size> layout=stripe ncol=3 <disk1> <disk2> <disk3>

Create mirrored volume with log vxassist make <volume> <size> layout=mirror, log nmirror=# nlog=#

Note: for information about logging see logging section below
Create a raid volume vxassist make <volume> <size> layout=raid5
Create a raw volume

vxassist -U <usage_type> make <volume> <size> alloc='<disk>'

types:
fsgen - filesystems
gen - raw volumes
raid - supports raid5
root - suuports root filesystems for booting
swap - performs no recovery on startup
relayout - used temporary for disk relayout operations

Remove a volume

vxedit rf rm <volume>
vxassist -g <group> remove volume <vol>

Note: you must disable the volume first

Initializing a volume vxvol init state <volume> [plex]state=clean,enable,active
Online Relayout

vxassist -g <group> relayout <vol> layout=stripe ncol=2
vxassist -g <group> relayout <vol> layout=stripe ncol=+1
vxassist -g <group> relayout <vol> layout=stripe ncol=-1
vxassist -g <group> relayout <vol> layout=stripe stripe=32k ncol=5
vxassist -g <group> relayout <vol> layout=raid5 stripeunit=32k ncol3
vxassist -g <group> convert <vol> layout=stripe-mirror

# Display the relayout operation
vxrelayout -g <group> [status|reverse|start] <vol>
vxtask list

Remove a volume off a particular disk ## either make sure there is a another disk in the group, you can even specify it
vxassist -g <group> move <vol> \!<disk> [<disk>]
Rename a volume vxedit -g <group> rename <old_vol_name> <new_vol_name>
Starting a volume vxvol start <volume>
Start a disabled volume

vxrecover sb <volume>

-s = start volume after recovery
-b = background the recovery task

Disable a volume vxvol -g <group> stop <volume>
Evacuate a volume vxevac -g <group> <from-disk> <to-disk>
Maintenance mode vxvol maint <volume>
Not clean vxmend mirror clean <plex>
No kernel state vxplex att <vol_name> <plex>
Extending a volume size vxresize <volume> <new length>
vxresize -g <group> <volume> +100m       - increase the volume by 100Mb
Shrinking a volume size vxresize <volume> <new length>
vxresize -g <group> <volume> -100m       - decrease the volume by 100Mb
add a DRL log to a volume vxassist addlog <volume>
remove a DRL log from a volume vxassist remove log <volume>
Extending log size vxvol set loglen = 2m <volume/log>
Detering volume size vxassit -g <group> maxsize layout=mirror - the maximum size you can create a mirror
vxassist -g <group> maxgrow <volume>     - the maximum size the volume can grow too
Recover a volume

Vxmend fix clean <plex>

Change volumes permissions vxedit -g <group> set owner=<user> group=<group> mode=<perms> <vol>

# to display the permissions
ls -l /dev/vx/rdsk/<group>

Plexs

Creating a plex vxmake plex <plex> sd =<sub disk name>
Remove a plex vxplex –o rm dis <plex>
vxplex -g <group> dis <plex>
vxedit -g <group> -rf rm <plex>
Moving a plex vxplex mv <original plex> <new plex>
Copying a plex vxplex cp <volume> <new plex>
Attaching a plex vxplex att <volume> <plex>
Detaching a plex vxplex det <plex>
Offlining a plex vxmend off vol01-02

Sub-disks

Creating sub-disk vxmake sd <sub-disk> <disk>, offset, len
Removing sub-disk vxedit rm <sub-disk>
Moving sub-disk vxsd mv <old sub-disk> <new sub-disk>
Associating with a plex vxmake plex <plex> sd=<sub-disk>, …
i.e vxmake plex home-1 sd= disk02-01, disk02-00, disk02-02
Dissociating vxsd dis <sub-disk>
Splitting vxsd –s<size> split sd<new sub><newsub2>
Joining vxsd join <sub-disk1><subdisk2><new subdisk>
relocating a sub disk ## either make sure there is a another disk in the group, you can even specify it
vxassist -g <group> move <vol> \!<disk> [<disk>]
relocating a whole disk sub disks

vxprint -g rootdg -se 'sd_orig_dmname="disk02"'
vxunreloc -g rootdg disk02


Volume Manager Information

Disks
Display all the physical disks

vxdisk list
vxdisk -o alldgs list

Display detailed disk info vxdisk list <disk>
Check for disk failures vxstat g <group>ff d
Disk Group
Display group properties vxdg list
Display detailed group info vxdg list <group>
  vxinfo -p -g <group>
Volume
Display volume info vxprint Aht <vol>
Display volume properties vxprint vl
Display unstartable volume vxinfo -g <group> <vol>
Check for volume failures vxstat g <group>ff -v
Plex
Display plex properties vxprint vp
Check for plex failures vxstat g <group>ff -p
Sub-Disks
Display sub-disk properties vxprint st
Veritas Tasks
Display tasks

vxtask list
vxtask monitor   - continuously monitor

States:
r = running
p = pause
a = aborting

Statistics and Tracing
Iostats vxstat -g <group> -r d <disks>       - reset all stats on disk
vxstat -g <group> -d                  - display stats
vxstat -g <group> -i 1 -d <vol>       - display stats every 1 sec intervals for volume
vxstat -g <group> -i 10 -c 5 -d       - display 5 sets @ 10 secs intervals
Tracing vxtrace -d <filename> -o dev,disk <vol>
vxtrace -f <filename> -o dev,disk <vol> | more

Licensing

Add vxlicinst               - versions greater than 3.5
vxlicense -c            - versions below 3.5
View vxlicrep                - versions greater than 3.5
vxlicense -p            - versions below 3.5
Paths

/etc/vx/licenses/lic    - versions greater than 3.5
/etc/vx/elm             - versions below 3.5

Reload new license vxdctl license init

VEA

Start/Stop /etc/init.d/isisd [start|stop|restart]
Status

vxsvcctrl status
vxsvc [-m|-k|-v]

-m = status
-k = kill
-v = version

Daemons /opt/VRTSob/bin/vxsvc
/opt/VRTSob/bin/vxsvcctrl
Start VEA GUI

/opt/VRTSob/bin/vea

You must have X-windows running

Logging

Logging help in recovery and can speed it up dramatically, the main form of logging in veritas is the DRL (dirty region log) which performs the following

Add vxassist -g <group> addlog <vol> logtype=drl
vxassist -g <group> addlog <vol>                       - used for raid logs (no type)
Remove vxassist -g <group> remove log <vol> [nlog=n] <vol>

Volume Read Policy

Policies can be used if you have slower disks within a volume and you wish to use the faster disks.

Round Robin vxvol -g <group> rdpol round <vol>
Preffered Plex vxvol -g <group> rdpol prefer <vol> <plex>
Selected Plex vxvol -g <group> rdpol select <vol>

Storage Expert (vxse)

Veritas have created some scripots that can check the integrity of the vxvm setup i.e mirrored volumes, spares, etc. The scripts are based on rules and there are a number of differents rules veritas has set, look in the rules directory to see all of them.

Display Description vxse_raid5log1 info
Check rules vxse_raid5log1 -g <group> check
List spare rules vxse_spares list
Run spare run

vxse_spares run

Note: you need to run "/etc/init.d/isisd start" to start the necessary daemons first

VXSE Paths
Rules /opt/VRTS/vxse/vxvm
Default Parameters /etc/default/vxse

VxDMP

see VxDMP for more information

Veritas links

Recovery features http://prefetch.net/articles/veritasrecoveryfeatures.html