Solstice DiskSuite 4.0 Administration Guide
  Rechercher uniquement dans ce livre
Télécharger cet ouvrage au format PDF

Introduction

1

The Solstice DiskSuite software product offers better performance, greater capacity, easier administration, and improved availability of disk storage on SPARC and x86 systems.
With Solstice DiskSuite, data availability and reliability are improved with three-way mirroring. You can mirror swap or any file system, including root and /usr. You can mirror existing file systems and automatically replace failed components within a mirror using hot spare facilities. Resynchronization time for mirrors is greatly reduced by optimized copying.
The UNIX file system (UFS) logging facility provides faster local directory operations, speeds up reboots, and decreases synchronous disk writes by recording file system updates in a log before they are applied to the UFS file system.
In addition, DiskSuite's disk striping and concatenation can increase performance. Striping spreads data requests over multiple components. Concatenation and stripes increase capacity by grouping several components into a single large logical device.
The information about the configuration and state of all metadevices is preserved in a replicated state database. This further ensures the integrity of data.
Administration is simplified by the hot spare facility and the dynamic growth of metadevices and file systems. New functionality included in the Solstice DiskSuite 4.0 release includes the creation of RAID (level 5) configurations and
disksets. A diskset is a grouping of two hosts and disk drives in which all the drives are accessible by both hosts. RAID provides recovery from disk failure in a more cost effective way than disk mirroring.
This introductory chapter provides the following:
  • Information on how to get help
  • A list of the peripherals supported by Solstice DiskSuite
  • A list of DiskSuite terms and their definitions

Getting Help

If you have problems installing or using Solstice DiskSuite, call the distributor from which you purchased the software and provide the following information:
  • Your name and electronic mail address (if available)
  • Your company name, address, and phone number
  • The model and serial number of your system
  • The release number of the operating system (for example, Solaris 2.3 or Solaris 2.4)
  • Any additional information that will help diagnose the problem

Supported Peripherals

Solstice DiskSuite 4.0 runs on all SPARC systems running Solaris 2.3 or a later Solaris 2.x release and on all x86 systems running Solaris 2.4 or a later Solaris 2.x release. (You must run Solaris 2.4 or a later Solaris 2.x release to use the UFS logging facility.)
The peripherals supported by DiskSuite include:
  • IPI disk drives
  • SCSI disk drives
  • IDE disk drives (x86 only)

Note - The root file system cannot be mirrored on an IDE drive.

Disk Naming

Backup Copilot uses the normal disk naming convention for all drives. This naming is, cntndnsn, where c is the controller number, t is the target, d is the disk, and s is the slice. However it is important to be aware that on x86 systems, IDE drives do not have a target number. All the examples and instructions in this manual are from a SPARC system and show the target number.

Definition of Terms

The following are general definitions for Solstice DiskSuite:
Attach submirror - to add a submirror to an existing metamirror by using the metattach(1M) command. The attached submirror is then resynced with other submirrors.
Attach logging device - to add a logging device to an existing metatrans device by using the metattach(1M) command.
Component - the physical partition that is part of a metadevice.
Concatenated Stripe - a metadevice that is made up of both concatenated and striped components.
Concatenation - the act of combining two or more physical components into a single metadevice. The partitions are accessed sequentially (treated as a single device) rather than interlaced (as with stripes).
Detach submirror - to remove a submirror from an existing metamirror by using the metadetach(1M) command.
Detach logging device - to remove a logging device from a metatrans device by using the metadetach(1M) command.
Diskset - a grouping of two hosts and disk drives in which all the drives are accessible by both hosts. This feature provides facilities for hosts to share disks in a high availabilty environment.
Hot spare - a component set up to be automatically substituted for a failed component of a mirrored or RAID metadevice.
Hot spare pool - a group of spare components which automatically replace failed components.
Interlace - the number of blocks on a component of a striped or RAID metadevice that can be accessed simultaneously with the same number of blocks from other components. The interlace value tells DiskSuite how much data to place on a component of a striped or RAID metadevice before moving on to the next component.
Local diskset - a diskset for a host consisting of all drives which are not part of a shared diskset.
Logging - recording UNIX file system (UFS) updates in a log (the logging device) before they are applied to the UNIX file system (the master device).
Logging device - the component of a metatrans device that contains the log.
Master device - the component of a metatrans device that contains the file system.
Metadevice - a group of components accessed as a single logical device through concatenating, striping, mirroring, logging the physical devices, or setting up RAID devices.
Metadevice state database - information kept in nonvolatile storage (on disk) for preserving the state and configuration of metadevices.
Metadriver - a pseudo device driver that maps operations on a metadevice into operations on its components.
Metamirror - a special type of metadevice (also referred to as a "Mirror"), composed of one or more other metadevices called submirrors.
Metatrans device - a special type of metadevice which implements UFS logging (also referred to as a "Trans device"). A metatrans device is composed of one or more other metadevices or components. This pseudo device is made up of a master device and, optionally, a logging device.
md.tab - the metadevice workspace file for metainit. Note that this file is maintained by the user.
Mirroring - replicating all writes to a single logical device (the metamirror) to multiple devices (the submirrors) while distributing read operations. This provides redundancy of data in the event of a failure.
Optimized resync - an update of only the submirror regions that are not in sync when the system reboots. Because DiskSuite keeps track of regions of all metamirrors, the metadisk driver knows which regions of the submirrors are not identical (in sync) on all submirrors after a crash.
Partial resync - resyncing only a replaced component of a submirror or RAID device, rather than the entire submirror or RAID device.
RAID - an acronym for Redundant Arrays of Inexpensive Disks.
Replica - a copy of the state database. Keeping copies of the state database protects against the loss of state and configuration information. This information is critical to the operation of all metadevices.
Resync region - a division of a metamirror that enables DiskSuite to track changes by region rather than over the entire metamirror. By dividing the metamirror into regions, resync time may be reduced.
Resyncing - copying data from one submirror to another after system crashes, submirror failures, or after using the metattach command to add a submirror. Resyncing ensures the data on all submirrors is identical.
State database - a dedicated portion of a disk reserved exclusively for the metadisk driver. It cannot be used for any other purpose.
Stripe - similar to concatenation, except the addressing of the component blocks is interlaced on the partitions, rather than addressed sequentially. Striping is used to gain performance. By striping data across disks, multiple controllers can access data simultaneously.
Submirror - a metadevice that is part of a metamirror.
UFS - an acronym for the UNIX file system.
UFS logging - recording UNIX file system updates to a log (the logging device) before the updates are applied to the UFS (the master device).