VERITAS Volume Manager

Databases require their storage media to be robust and resilient to failure. It is vital to protect against hardware and disk failures and to maximize performance using all the available hardware resources. Using a volume manager provides this necessary resilience and eases the task of management. A volume manager can help you manage hundreds of disk devices and makes spanning, striping, and mirroring easy.

VERITAS Volume Manager (VxVM) builds virtual devices called volumes on top of physical disks. Volumes are accessed by a file system, a database, or other applications in the same way physical disk partitions would be accessed. Using volumes, VxVM provides the following administrative benefits for databases:

• Spanning of multiple disks---eliminates media size limitations.
• Striping---increases throughput and bandwidth.
• Mirroring or RAID-5---increases data availability.
• Online relayout---allows online volume layout changes to improve database performance.
• Volume resynchronization---ensures that all mirrors contain exactly the same data and that the data and parity in RAID-5 volumes agree.
• Dirty Region Logging (DRL)---speeds the recovery of mirrored volumes after a system crash.
• SmartSync Recovery Accelerator---increases the availability of mirrored volumes by only resynchronizing changed data.
• Volume snapshots---allows backup of volumes based on disk mirroring. With this release, VxVM introduces full-sized and space-optimized instant snapshots, the preferred way to implement online and off-host point-in-time copy solutions.
• FastResync---separately licensed, optional feature that performs quick and efficient resynchronization of stale mirrors. FastResync is included with the Enterprise Edition and is also included as part of the VERITAS FlashSnap option with the Standard Edition.
• Disk group split and join---separately licensed, optional feature that supports general disk group reorganization and allows you to move volume snapshots to another host for off-host backup. Disk group split and join is included with the Enterprise Edition and is also included as part of the VERITAS FlashSnap option with the Standard Edition.
• Hot-relocation---automatically restores data redundancy in mirrored and RAID-5 volumes when a disk fails.
• Dynamic multipathing (DMP)---allows for transparent failover, load sharing, and hot plugging of SCSI devices.
• Volume sets---allows several volumes to be represented by a single logical mount device.
• Dynamic LUN Expansion---allows you to resize a disk after it has been initialized while preserving the existing data on the disk.
• Storage Expert---helps diagnose configuration problems with VxVM.
• Cluster Volume Manager (CVM)---separately licensed, optional feature that allows you to use VxVM in a cluster environment.
• VERITAS FlashSnap Agent for Symmetrix---separately licensed, optional feature that includes a set of commands that allows you to use EMC TimeFinder in conjunction with VxVM disk groups and volumes that have been created on Symmetrix standard devices.
• VERITAS Volume Replicator (VVR)---separately licensed, optional feature that provides data replication for disaster recovery planning.
• Free space pool management---simplifies administration and provides flexible use of available hardware.
• Online administration---allows configuration changes without system or database down time.

The following sections provide brief overviews of VxVM concepts and features that are relevant to database administration. The information and examples presented in the remainder of this guide assume that you are using VERITAS Volume Manager. For a more detailed description of VxVM and its features, refer to the VERITAS Volume Manager Administrator's Guide.

Volumes

A volume is a virtual disk device that appears to applications, databases, and file systems like a physical disk partition without the physical limitations of a disk partition. A volume consists of one or more plexes, each holding a copy of the selected data in the volume. Due to its virtual nature, a volume is not restricted to a particular disk or a specific area of a disk. For example, a volume can span multiple disks and can be used to create a large file system.

Volumes consist of other virtual objects that can be manipulated to change the volume's configuration. Volumes and their virtual components are referred to as Volume Manager objects. You can manipulate VERITAS Volume Manager objects in a variety of ways to optimize performance, provide redundancy of data, and perform backups or other administrative tasks on one or more physical disks without interrupting applications. As a result, data availability and disk subsystem throughput are improved.

You can change the configuration of a volume without causing disruption to databases or file systems that are using the volume. For example, you can mirror a volume on separate disks or move the volume to use different disk storage.

Disk Groups

A disk group is a collection of disks that share a common configuration (for example, configuration objects that belong to a single database). We recommend creating one disk group for each database.

You can move a disk group and its components as a unit from one host to another host. For example, you can move volumes and file systems that belong to the same database and are created within one disk group as a unit. You must configure a given volume from disks belonging to one disk group.

In releases before VERITAS Storage Foundation 4.0 for Oracle, the default disk group was rootdg. For VxVM to function, the rootdg disk group had to exist and it had to contain at least one disk. This requirement no longer exists, and VxVM can work without any disk groups configured (although you must set up at least one disk group before you can create any volumes of other VxVM objects).

Volume Layouts

A Redundant Array of Independent Disks (RAID) is a disk array in which a group of disks appears to the system as a single virtual disk or a single volume. VxVM supports several RAID implementations, as well as spanning. The following volume layouts are available to satisfy different database configuration requirements:

• Spanning and concatenation
• Striping (RAID-0)
• Mirroring (RAID-1)
• Mirrored-Stripe Volumes (RAID-0+1)
• Striped-Mirror Volumes (RAID-1+0)
• RAID-5

---

 Caution    Spanning or striping a volume across multiple disks increases the chance that a disk failure will result in failure of that volume. Use mirroring or RAID-5 to substantially reduce the chance of a single volume failure caused by a single disk failure.

---

Spanning and Concatenation

Concatenation maps data in a linear manner onto one or more subdisks in a plex. To access all of the data in a concatenated plex sequentially, data is first accessed in the first subdisk from beginning to end. Data is then accessed in the remaining subdisks sequentially from beginning to end, until the end of the last subdisk.

The subdisks in a concatenated plex do not have to be physically contiguous and can belong to more than one VM disk. Concatenation using subdisks that reside on more than one VM disk is called spanning.

Spanning is useful when you need to read or write data sequentially (for example, reading from or writing to database redo logs) and there is not sufficient contiguous space.

Striping (RAID-0)

Striping is a technique of mapping data so that the data is interleaved among multiple physical disks. Data is allocated in equal-sized units (called stripe units) that are interleaved between the disks. Each stripe unit is a set of contiguous blocks on a disk. A stripe consists of the set of stripe units at the same position across all columns. A column is a set of one or more subdisks within a striped plex.

Striping is useful if you need large amounts of data written to or read from physical disks, and performance is important. Striping is also helpful in balancing the I/O load from multi-user applications across multiple disks. By using parallel data transfer to and from multiple disks, striping significantly improves data-access performance.

When striping across multiple disks, failure of any one disk will make the entire volume unusable.

Mirroring (RAID-1)

Mirroring is a technique of using multiple copies of the data, or mirrors, to duplicate the information contained in a volume. In the event of a physical disk failure, the mirror on the failed disk becomes unavailable, but the system continues to operate using the unaffected mirrors. For this reason, mirroring increases system reliability and availability. A volume requires at least two mirrors to provide redundancy of data. A volume can consist of up to 32 mirrors. Each of these mirrors must contain disk space from different disks for the redundancy to be effective.

Striping Plus Mirroring (Mirrored-Stripe or RAID-0+1)

VxVM supports the combination of mirroring with striping. When used together on the same volume, mirroring plus striping offers the benefits of spreading data across multiple disks while providing redundancy of data.

Mirrored-stripe volumes have multiple plexes as mirrors, each constructed as a striped plex. Allocate subdisks used in the same striped plex from separate disks, and use a disk in only one mirror of a volume.

Mirroring Plus Striping (Striped-Mirror Volumes, RAID-1+0 or RAID-10)

VxVM supports the combination of striping with mirroring. When used together on the same volume, striping plus mirroring offers the benefits of spreading data across multiple disks while providing redundancy of data.

Striped-mirror volumes combine striping and mirroring, but the mirroring is done at stripe column level. In case of failure, this type of volume recovers faster than RAID-0+1 volumes and the tolerance for disk failure is greater.

For databases that support online transaction processing (OLTP) workloads, we recommend either mirrored-stripe or striped-mirror volumes to improve database performance and reliability. For highest availability, we recommend striped-mirror volumes (RAID 1+0).

RAID-5 (Striping with Parity)

RAID-5 provides data redundancy through the use of parity (a calculated value that the system uses to reconstruct data after a failure). While data is written to a RAID-5 volume, parity is also calculated by performing an exclusive OR (XOR) procedure on data. The resulting parity is then written to another part of the volume. If a portion of a RAID-5 volume fails, the data that was on that portion of the failed volume can be recreated from the remaining data and the parity.

RAID-5 offers data redundancy similar to mirroring, while requiring less disk space. RAID-5 read performance is similar to that of striping but with relatively slow write performance. RAID-5 is useful if the database workload is read-intensive (as in many data warehousing applications). You can snapshot a RAID-5 volume and move a RAID-5 subdisk without losing redundancy.

Online Relayout

As databases grow and usage patterns change, online relayout lets you change volumes to a different layout, with uninterrupted data access. Relayout is accomplished online and in place. Use online relayout to change the redundancy or performance characteristics of the storage, such as data organization (RAID levels), the number of columns for RAID-5 and striped volumes, and stripe unit size.

Volume Resynchronization

When storing data redundantly, using mirrored or RAID-5 volumes, VERITAS Volume Manager ensures that all copies of the data match exactly. However, if the system crashes, small amounts of the redundant data on a volume can become inconsistent or unsynchronized. For mirrored volumes, unsynchronized data can cause two reads from the same region of the volume to return different results if different mirrors are used to satisfy the read request. In the case of RAID-5 volumes, unsynchronized data can lead to parity corruption and incorrect data reconstruction.

In the event of a system crash, VERITAS Volume Manager ensures that all mirrors contain exactly the same data and that the data and parity in RAID-5 volumes agree. This process is called volume resynchronization. Not all volumes require resynchronization after a system failure. VxVM notices when a volume is first written and marks it as dirty. Only volumes that are marked dirty when the system reboots require resynchronization.

The process of resynchronization can impact system and database performance. However, it does not affect the availability of the database after system reboot. You can immediately access the database after database recovery although the performance may suffer due to resynchronization. For very large volumes or for a very large number of volumes, the resynchronization process can take a long time. You can significantly reduce resynchronization time by using Dirty Region Logging (DRL) for mirrored volumes or by making sure that RAID-5 volumes have valid RAID-5 logs. However, using logs can slightly reduce the database write performance.

For most database configurations, we recommend using dirty region logs or the RAID-5 logs when mirrored or RAID-5 volumes are used. It is also advisable to evaluate the database performance requirements to determine the optimal volume configurations for the databases.

Dirty Region Logging

Dirty Region Logging (DRL), if enabled, speeds the recovery of mirrored volumes after a system crash. DRL keeps track of the regions that have changed due to I/O writes to a mirrored volume. DRL uses this information to recover only those portions of the volume that need to be recovered.

SmartSync Recovery Accelerator

SmartSync increases the availability of mirrored volumes by only resynchronizing changed data. SmartSync reduces the time required to restore consistency, freeing more I/O bandwidth for business-critical applications.

This feature is applicable only to databases that are configured on raw volumes. If supported by the database vendor, the SmartSync feature uses an extended interface between VxVM volumes and the database software to avoid unnecessary work during mirror resynchronization. For example, Oracle automatically takes advantage of SmartSync to perform database resynchronization when it is available.

Volume Sets

Volume sets are an enhancement to VxVM that allow several volumes to be represented by a single logical mount device. All I/O from and to the underlying volumes is directed via the I/O interfaces of the volume set. The volume set feature supports the multi-device enhancement to VERITAS File System (VxFS). This feature allows file systems to make best use of the different performance and availability characteristics of the underlying volumes. For example, file system metadata could be stored on volumes with higher redundancy, and user data on volumes with better performance.

Volume Snapshots

A volume snapshot is a point-in-time image of a volume. VERITAS Volume Manager provides three volume snapshot features based on disk mirroring:

• Full-sized instant snapshots
• Space-optimized instant snapshots
• Emulation of third-mirror snapshots

For detailed information on performing database backups using volume snapshots, see Using Database FlashSnap for Backup and Off-Host Processing

VERITAS FastResync (Optional)

VERITAS FastResync (previously called Fast Mirror Resynchronization or FMR) is included with the Enterprise Edition. It is also included as part of the VERITAS FlashSnap option with the Standard Edition.

VERITAS FastResync performs quick and efficient resynchronization of stale mirrors (mirrors that are not synchronized). This increases the efficiency of the VxVM snapshot mechanism, and improves the performance of operations such as backup and decision support. Typically, these operations require that the volume is quiescent, and that they are not impeded by updates to the volume by other activities on the system. To achieve these goals, the snapshot mechanism in VxVM creates an exact copy of a primary volume at an instant in time. After a snapshot is taken, it can be accessed independently of the volume from which it was taken.

VERITAS Storage Foundation for Oracle Enterprise Edition includes a feature called Database FlashSnap, which takes advantage of the FastResync and disk group split and join features. Database FlashSnap provides a quicker and easier way for database administrators to use volume snapshots. For more information, see VERITAS Database FlashSnap.

Non-Persistent FastResync

Non-persistent FastResync allocates its change maps in memory. If non-persistent FastResync is enabled, a separate FastResync map is kept for the original volume and for each snapshot volume. Unlike a dirty region log (DRL), these maps do not reside on disk nor in persistent store. The advantage is that updates to the FastResync map have little impact on I/O performance, as no disk updates need to be performed. However, if a system is rebooted, the information in the map is lost, so a full resynchronization is required when performing a snapback operation. This limitation can be overcome for volumes in cluster-shareable disk groups, provided that at least one of the nodes in the cluster remains running to preserve the FastResync map in its memory.

Persistent FastResync

Non-persistent FastResync has been augmented by the introduction of persistent FastResync. Unlike non-persistent FastResync, Persistent FastResync keeps the FastResync maps on disk so that they can survive system reboots and system crashes. When the disk groups are rejoined, this allows the snapshot plexes to be quickly resynchronized. This ability is not supported by non-persistent FastResync.

If persistent FastResync is enabled on a volume or on a snapshot volume, a DCO and a DCO log volume are associated with the volume.

In VxVM 3.2 and 3.5, the DCO object only managed information about the FastResync maps. These maps track writes to the original volume (and to each of up to 32 snapshot volumes) since the last snapshot operation. The DCO log volume on disk holds the 33 maps, each of which is 4 blocks in size by default.

In VxVM 4.1, the DCO object is used not only to manage FastResync maps, but also to manage DRL recovery maps and special maps called copy maps that allow instant snapshot operations to be resume following a system crash.

Persistent FastResync can also track the association between volumes and their snapshot
volumes after they are moved into different disk groups. When the disk groups are rejoined, this allows the snapshot plexes to be quickly resynchronized. This ability is not supported by non-persistent FastResync.

Disk Group Split and Join (Optional)

Disk group split and join is included with the Enterprise Edition. It is also included as part of the VERITAS FlashSnap option with the Standard Edition.

VxVM provides a disk group content reorganization feature that supports general disk group reorganization and allows you to move volume snapshots to another host for off-host backup. Additional options to the vxdg command enable you to take advantage of the ability to remove all VxVM objects from an imported disk group and move them to a newly created target disk group (split), and to remove all VxVM objects from an imported disk group and move them to an imported target disk group (join). The move operation enables you to move a self-contained set of VxVM objects between the imported disk groups.

Hot-Relocation

In addition to providing volume layouts that help improve database performance and availability, VxVM offers features that you can use to further improve system availability in the event of a disk failure. Hot-relocation is a feature that allows a system to react automatically to I/O failures on mirrored or RAID-5 volumes and restore redundancy and access to those volumes.

VxVM detects I/O failures on volumes and relocates the affected portions to disks designated as spare disks or free space within the disk group. VxVM then reconstructs the volumes that existed before the failure and makes them redundant and accessible again.

The hot-relocation feature is enabled by default and is recommended for most database configurations. After hot-relocation occurs, we recommend verifying the volume configuration for any possible performance impact. It is also a good idea to designate additional disks as spares to augment the spare pool.

While a disk is designated as a spare, you cannot use the space on that disk for the creation of VxVM objects within its disk group. VxVM also lets you free a spare disk for general use by removing it from the pool of hot-relocation disks.

DMP-Supported Disk Arrays

VxVM provides administrative utilities and driver support for disk arrays that can take advantage of its Dynamic Multipathing (DMP) feature. Some disk arrays provide multiple ports to access their disk devices. These ports, coupled with the host bus adaptor (HBA) controller and any data bus or I/O processor local to the array, make up multiple hardware paths to access the disk devices. Such disk arrays are called multipathed disk arrays. This type of disk array can be connected to host systems in many different configurations, (such as multiple ports connected to different controllers on a single host, chaining of the ports through a single controller on a host, or ports connected to different hosts simultaneously). DMP is available for multiported disk arrays from various vendors and provides improved reliability and performance by using path failover and load balancing.

See the VERITAS Volume Manager Administrator's Guide for detailed information and the VERITAS Volume Manager Hardware Notes for information about supported disk arrays.

Dynamic LUN Expansion

Dynamic LUN expansion allows you to resize a disk after it has been initialized while preserving the existing data on the disk. See the VERITAS Volume Manager Administrator's Guide for more information.

Storage Expert

Storage Expert consists of a set of simple commands that collect VxVM configuration data and compare it with "best practice." Storage Expert then produces a summary report that shows which objects do not meet these criteria and makes recommendations for VxVM configuration improvements.

These user-configurable tools help you as an administrator to verify and validate systems and non-optimal configurations in both small and large VxVM installations.

Storage Expert components include a set of rule scripts and a rules engine. The rules engine runs the scripts and produces ASCII output, which is organized and archived by Storage Expert's report generator. This output contains information about areas of VxVM configuration that do not meet the set criteria. By default, output is sent to the screen, but you can redirect it to a file using standard UNIX redirection. See the VERITAS Volume Manager Administrator's Guide for more information.

Cluster Functionality (Optional)

VxVM includes an optional, separately licensable clustering feature, known as Cluster Volume Manager, that enables VxVM to be used in a cluster environment. With the clustering option, VxVM supports up to 16 nodes per cluster. See the VERITAS Volume Manager Administrator's Guide for more information.

VERITAS Volume Replicator (Optional)

VERITAS Volume Replicator (VVR) is an optional, separately licensable feature of VxVM. VVR is a data replication tool designed to maintain a consistent copy of application data at a remote site. It is built to contribute to an effective disaster recovery plan. If the data center is destroyed, the application data is immediately available at the remote site, and the application can be restarted at the remote site.

VVR works as a fully integrated component of VxVM. VVR benefits from the robustness, ease of use, and high performance of VxVM and, at the same time, adds replication capability to VxVM. VVR can use existing VxVM configurations with some restrictions. Any application, even with existing data, can be configured to use VVR transparently.

See the VERITAS Volume Replicator documentation for more information.