Beginners Lesson – VERITAS Volume Manager for Solaris
发布时间:2016-02-17 23:18:09
发布时间:2016-02-17 23:18:09
Beginner's Lesson – VERITAS Volume Manager for Solaris
June 4, 2011
By unixadminschool com
Veritas Volume Manager is a storage management application by Symantec , which allows you to manage physical disks as logical devices called volumes.
VxVM uses two types of objects to perform the storage management
1. Physical objects - are direct mappings to physical disks
2 . Virtual objects - are volumes, plexes, subdisks and diskgroups.
a. Disk groups are composed of Volumesb. Volumes are composed of Plexes and Subdisksc. Plexes are composed of SubDisksd. Subdisks are actual disk space segments of VxVM disk ( directly mapped from the physical disks)
1. Physical Disks
Physical disk is a basic storage where ultimate data will be stored. In Solaris physical disk names uses the convention like “c#t#d#” where c# refers to controller/adapter connection, t# refers to the SCSI target Id , and d# refers to disk device Id. Below figure illustrates how the disk name changes depending on the connection.
Physical disks could be coming from different sources within the servers e.g. Internal disks to the server , Disks from the Disk and Disks from the SAN.
Check the disks recognized by Solaris
#echo|formatSearching for disks…done
AVAILABLE DISK SELECTIONS:0. c0t0d0 /sbus@1f,0/SUNW,fas@e,8800000/sd@0,01. c0t1d0 /sbus@1f,0/SUNW,fas@e,8800000/sd@1,0
2. Solaris Native Disk Partitioning
In solaris, physical disks will partitioned into slices numbered as S0,S1,S3,S4,S5,S6,S7 and the slice number S2 normally called as overlap slice and points to the entire disk. In Solaris we use the format utility used to partition the physical disks into slices.
Once we added new disks to the Server, first we should recognize the disks from the solaris level before proceeding for any other storage management utility.
Steps to add new disk to Solaris:
If the disks that are recently added to the server not visible, you can use below procedure
Option 1: Reconfiguration Reboot ( for the server hardware models that doesn’t support hot swapping/dynamic addition of disks )
# touch /reconfigure; init 6
or
#reboot — -r ( only if no applications running on the machine)
Option 2: Recognize the disks added to external SCSI, without reboot
# devfsadm
# echo | format
Option 3: Recognize disks that are added to internal scsi, hot swappable, disk connections.
Just run the command “cfgadm -al” and check for any newly added devices in “unconfigured” state, and configure them.
# cfgadm -alAp_Id Type Receptacle Occupant Conditionc0 scsi-bus connected configured unknownc0::dsk/c0t0d0 disk connected configured unknownc0::dsk/c0t0d0 disk connected configured unknownc0::rmt/0 tape connected configured unknownc1 scsi-bus connected configured unknownc1::dsk/c1t0d0 unavailable connected unconfigured unknown c1::dsk/c1t1d0 unavailable connected unconfigured unknown < == disk not configured
# cfgadm -c configure c1::dsk/c1t0d0
# cfgadm -c configure c1::dsk/c1t0d0
# cfgadm -alAp_Id Type Receptacle Occupant Conditionc0 scsi-bus connected configured unknownc0::dsk/c0t0d0 disk connected configured unknownc0::rmt/0 tape connected configured unknownc1 scsi-bus connected configured unknownc1::dsk/c1t0d0 disk connected configured unknown c1::dsk/c1t1d0 disk connected configured unknown
# devfsadm
#echo|format
Additional Procedure : Click here to check the procedure to add Internal FC-AL disks
3. Initialize Physical Disks under VxVM control
A formatted physical disk is considered uninitialized until it is initialized for use by VxVM. When a disk is initialized, partitions for the public and private regions are created, VM disk header information is written to the private region and actual data is written to Public region. During the notmal initialization process any data or partitions that may have existed on the disk are removed.
Note: Encapsulation is another method of placing a disk under VxVM control in which existing data on the disk is preserved
An initialized disk is placed into the VxVM free disk pool. The VxVM free disk pool contains disks that have been initialized but that have not yet been assigned to a disk group. These disks are under Volume Manager control but cannot be used by Volume Manager until they are added to a disk group
Device Naming Schemes
In VxVM, device names can be represented in two ways:
∙ Using the traditional operating system-dependent format c#t#d#
∙ Using an operating system-independent format that is based on enclosure names
c#t#d# Naming Scheme
Traditionally, device names in VxVM have been represented in the way that the operating system represents them. For example, Solaris and HP-UX both use the format c#t#d# in device naming, which is derived from the controller, target, and disk number. In VxVM version 3.1.1 and earlier, all disks are named using the c#t#d# format. VxVM parses disk names in this format to retrieve connectivity information for disks.
Enclosure-Based Naming Scheme
With VxVM version 3.2 and later, VxVM provides a new device naming scheme, called enclosure-based naming. With enclosure-based naming, the name of a disk is based on the logical name of the enclosure, or disk array, in which the disk resides.
Steps to Recognize new disks under VxVM control
1. Run the below command to see the available disks under VxVM control
# vxdisk list
in the output you will see below status
∙ error indicates that the disk has neither been initialized nor encapsulated by VxVM. The disk is uninitialized.
∙ online indicates that the drive has been initialized or encapsulated.
∙ online invalid indicated that disk is visible to VxVM but not controlled by VxVM
If disks are visible with “format” command but not visible with “vxdisk list” command, run below command to scan the new disks for VxVM
# vxdctl enable
Now you should see new disks with the status of “Online Invalid“
2. Initialize each disk with “vxdisksetup” command
#/etc/vx/bin/vxdisksetup -i
after running this command “vxdisk list” should see the status as “online” for all the newly initialized disks
4. Virtual Objects (DiskGroups / Volumes / Plexs ) in VxVM
Disk Groups
A disk group is a collection of VxVM disks ( going forward we will call them as VM Disks ) that share a common configuration. Disk groups allow you to group disks into logical group of Subdisks called plexes which in turn forms the volumes.
Volumes
A volume is a virtual disk device that appears to applications, databases, and file systems like a physical disk device, but does not have the physical limitations of a physical disk device. A volume consists of one or more plexes, each holding a copy of the selected data in the volume.
Plexes:
VxVM uses subdisks to create virtual objects called plexes. A plex consists of one or more subdisks located on one or more physical disks.
From the below diagram you can observer below points
∙ The Diskgroup named “Diskgroup_o1” is created using 4 different VM disks named as “vxdisk_0x, vxdisk_oy, vxdisk_oz and vxdisk_oa“
∙ The diskfgroup “Diskgroup_01″ was configured with 4 different volumes i.e. “1. Concat_vol 2. striped_vol 3. mirror_vol and 4. raid5_vol “
∙ Concat_vol is a concatenation volume with single plex i.e. con_plex01 inside, and con_plex01 was build up using 4 subdisks of different size
∙ Striped_vol is a striped volume with single plex i.e. stripe_plex01 inside, and stripe_plex01 was build up using 4 subdisks of same size
∙ mirror_vol is a mirrored volume with two plexes named “mplex01 and mplex02” inside, each plex is copy of other. Both the plexes formed with different subdisks of either same size or different size.
∙ raid5_vol is a raid5 ( striped with parity) volume which build up using 3 different plexes formed with the subdisks of 3 different VM disks.
Summary – Transformation of Physical disks into Veritas Volumes
Below diagram shows you the complete transformation of a physical disk into a veritas volume. And below is the summary of complete process
1. Recognize disks under solaris using devfsadm, cfgadm or reconfiguration reboot , and verify using format command2. Recognize the disks under VxVM using “vxdctl enable“3. Initialize the disks under VxVM using vxdisksetup4. Add the disks to Veritas Disk Group using vxdg commands5. Create Volumes under Disk Group using vxmake or vxassist commands6. Create filesystem on top of volumes using mkfs or newfs, and you can create either VXFS filesystem or UFS filesystem
Solaris Volume Manager Vs Veritas Volume Manager
Below diagram explains how Solaris volume manager differs from the Veritas volume manager during the process of creating a new file system on top of physical storage.