Volume OverviewDell OpenManage™ Array Manager 3.4
Dynamic Disk and Volume Support on NetWare, Windows Server 2003, and Linux
This chapter describes conceptual and procedural information about how Array Manager implements basic and dynamic volumes.
 |
Note There are particular considerations regarding dynamic disks and volumes on NetWare, Windows Server 2003, and Linux. See "Dynamic Disk and Volume Support on NetWare, Windows Server 2003, and Linux" for more information. |
A volume is a logical entity that is made up of portions of one or more physical disks. A volume can be formatted with a file system and can be accessed by a drive letter.
Like disks, volumes can be basic or dynamic. In Array Manager, basic volumes refer to all volumes that are not on dynamic disks. These include partitions and volumes managed by Array Manager as well as volumes created with Microsoft® Windows NT® Disk Administrator 4.0 or earlier. You cannot use the Array Manager and Disk Administrator applications at the same time. You can access volumes and partitions created by both applications simultaneously. However, if you are using a Windows NT 4.0 machine, you should continue to use Disk Administrator to change the volumes it created. See "Manage Volumes Created with Windows NT Disk Administrator" in this chapter for more information.
Dynamic volumes are logical volumes created from dynamic disks with Array Manager. For more information, see "Working with Dynamic Volumes" in this section.
There are particular considerations regarding dynamic disks and volumes on NetWare, Windows Server 2003, and Linux. See "Dynamic Disk and Volume Support on NetWare, Windows Server 2003, and Linux" for more information.
The remainder of this chapter describes volume functions and has three main topics:
Just as basic and dynamic disks share some common disk procedures, basic and dynamic volumes also share common volume procedures. Those procedures are documented in this section. The section's topics include:
A drive path refers to the path to an NTFS volume that is mounted at an empty folder. Mounting a volume at an empty folder is a feature of the Windows® 2000 operating system. It is not available in Windows NT 4.0. On Windows 2000, the Volume context menu contains the command Change Drive Letter and Path; on Windows NT 4.0, the corresponding command is Change Drive Letter.
The Change Drive Letter and Path or Change Drive Letter function is not available when formatting or upgrading to dynamic any disks contained in the volume. The format or upgrade to dynamic operation must complete before you can change the drive letter or path.
Note that you cannot reassign the drive letter of your active volume, the volume that contains your boot disk. You also cannot mount a volume on your boot disk.
|
Caution Changing the drive letter or drive path on a partition or volume may cause programs installed on that partition or volume to no longer run. |
The Change Drive Letter window displays.
The Drive Letter and Paths screen appears.
If a partition or volume does not have a drive letter, click Add. If you want to change an existing drive letter, click Modify. If you want to remove a drive letter or drive path, click Remove.
The Add Drive Letter and Path screen appears.
If a drive letter is already assigned, you will not be able to add another drive letter. The option will be inactive (dimmed).
You can add multiple drive paths to a partition or volume.
For details on mounting a volume at an empty folder, see "Mount a Volume at an Empty Folder" in the next section of this chapter.
The Modify Drive and Path screen appears.
If you try to assign a drive letter to a volume that is in use, a dialog box will ask you whether you want to forcibly reassign the drive letter. You can click Yes to do so.
If you want to modify an existing drive path, you will not be able to do so in this screen. You have to remove the drive path and recreate it. Click Cancel to return to the first screen and click Remove. Then you have to invoke the command again to recreate the drive path with the Add button.
You can add multiple drive paths to a partition or volume.
For details on mounting a volume at an empty folder, see "Mount a Volume at an Empty Folder" in the next section of this chapter.
Array Manager allows you to mount a volume at any empty folder on a local NTFS volume if you are using Windows 2000. This capability is not supported in Windows NT 4.0.
Array Manager assigns a drive path to the volume rather than a drive letter. Drive paths are useful because they eliminate the 24-drive-letter limit on hard-disk volume names. (Drive letters A and B are reserved for floppy drives.) The volume can be a partition, a logical drive that was created in Disk Administrator, or a dynamic volume.
The only restrictions are that the folder you mount the volume to must be empty and it must be on an NTFS volume on a local computer. You cannot mount volumes to network-accessible drives. You can, however, assign multiple drive paths to a single volume.
Mounted drive folders are distinguished from ordinary folders by having a drive icon instead of the normal folder icon.
You can assign the drive path when you create the volume with the Create Volume wizard at the Assign Drive Letter screen. You can also use the command Change Drive Letter and Path after the volume is created. A volume that already has a drive letter assigned can also have one or more drive paths. It does not have to be an unlabeled volume.
If you have not already created the volume, use the Create Volume wizard to create the volume. Do the following:
If you have not already created the folder, you can click the New Folder button to have Array Manager create the folder for you.
To unmount a volume with a drive path, use the command Change Drive Letter and Path, select the drive path, and click Remove. The capability of mounting and unmounting a volume at an empty folder is supported only in Windows 2000. It is not supported in Windows NT 4.0.
|
Note To modify a drive path, remove it, and then create a new drive path using the new location. You cannot modify the drive path directly. |
This command lets you view all the mounted volumes on a selected server. These volumes have been assigned a drive path rather than a drive letter. The capability of mounting a volume at a folder is supported by the Windows 2000 operating system. It is not supported in Windows NT 4.0.
The View All Drive Paths window appears.
You can also use the Unmount command or the Change Drive Letter and Path command to remove a drive path.
See also "Mount a Volume at an Empty Folder" in this section.
The Properties window comes up.
|
|
Caution All data in the volume or partition will be lost when you format it. |
Array Manager allows you to format using either FAT or NTFS. In addition, for Windows 2000 only, you can use FAT32.
|
Note It is strongly recommended that you leave the allocation unit size as Default. Selecting any other value requires an advanced knowledge of Windows file systems. You should thoroughly review the Microsoft Windows documentation before selecting any value other than Default. |
|
Note You can cancel formatting at any time during the process by selecting Cancel format from the menu. However, the Cancel option does not attempt to restore any of the volume's content. |
To refresh, click View and then Refresh. Progress bars display the percentage of progress while this command runs.
When you click the Refresh command, Array Manager updates displayed data with the current information. Refresh also checks to see whether unreadable volumes are now readable. However, it does not scan hardware. To update hardware information, use Rescan instead.
|
|
Caution All data in the volume or partition will be lost when you delete it. Make sure that the information has been backed up onto another storage medium and verified, or that it is no longer needed. |
|
Note You cannot delete the volume with the system files (the boot volume). |
The functions that apply only to basic volumes or partitions include:
The first two functions are covered in the "Disk Management" chapter. The remainder are covered in this section.
You can change the partition your system uses to start up by marking a selected primary partition as active.
The names commonly used for partitions containing the startup and operating system files are the system and boot partitions, respectively. The system partition refers to the partition containing the startup files. The boot partition contains the operating system files.
The system partition must be a primary partition that has been marked as active for startup purposes and must be located on the disk that the computer accesses when starting up the system. There can be only one active system partition at a time, which displays as Active in the status field. If you want to use another operating system, you must first mark its system partition as active before restarting the computer.
|
Note You cannot change the drive letter of an active partition. |
The utility partition is a factory-installed, bootable partition on the hard-disk drive that provides utilities and diagnostics for your Dell PowerEdge™ system. When activated, the partition boots and provides an executable environment for the partition's utilities. When the utility partition is not booted, it is designated as a non-MS-DOS partition.
This partition is placed on your hard-disk drive at the factory; it is recommended that you leave it as is.
Array Manager does not allow you to change Disk Administrator volumes, but it lets you view and do limited management of them. If you are running on Windows NT 4.0, you will need to exit Array Manager and then bring up Disk Administrator to make changes to these legacy Disk Administrator volumes.
|
|
Caution Disk Administrator should never be used while Array Manager is running, especially if there are tasks running on the controller at the time. Data loss can occur if both applications are running simultaneously. |
If you are using Windows 2000, you obviously will not be able to use Disk Administrator to change these volumes, but you can upgrade these volumes to dynamic volumes. See the next topic, "Upgrade Disks with Legacy Volumes to Dynamic," for details.
Array Manager recognizes volumes created with Windows NT Disk Administrator, but provides only a limited set of functions to use to administer these "legacy" volumes.
Array Manager uses different names to refer to the Disk Administrator volumes. The following table shows the respective names and the functions that Array Manager can perform on these volumes:
Upgrading a disk to dynamic that contains basic partitions and/or the mirrored, striped, spanned, or RAID-5 volumes configured with Disk Administrator cannot be done directly under Windows NT 4.0, but this upgrade can be done under Windows 2000.
There are particular considerations regarding dynamic disks and volumes on NetWare, Windows Server 2003, and Linux. See "Dynamic Disk and Volume Support on NetWare, Windows Server 2003, and Linux" for more information.
Under Windows NT 4.0, if you want to convert basic partitions to simple dynamic volumes or to convert Disk Administrator RAID volumes to dynamic volumes of the same type, you will have to back up your data, use Disk Administrator to delete the partitions or volumes, and exit Disk Administrator. Then you can start Array Manager, upgrade the disk or disks that contain the partitions or volumes to dynamic, and create new dynamic volumes of the types desired in Array Manager. Then you would need to copy back your data to the new dynamic volumes.
Under Windows 2000, you can upgrade partitions and NT 4.0 Disk Administrator RAID volumes to dynamic volumes by upgrading the disk(s) they are on to dynamic. This feature is called encapsulation.
With Windows 2000, you can upgrade basic disks containing these volumes to dynamic disks, and Array Manager will convert the partitions into volumes of the appropriate type, preserving all data contained in them. This upgrade can be done on any basic disk, including those containing the system and/or boot partitions.
If you are upgrading a Disk Administrator basic volume that resides on multiple disks, you must upgrade all the basic disks that contain the multi-disk volume. You must upgrade all of these disks together. Each of the disks must have at least 1MB of unallocated space at the end of the disk or the upgrade will fail. This space is needed for the database information that Array Manager uses to administer the volumes.
When you upgrade disks that contain partitions and NT Disk Administrator basic volumes, any existing partitions on the basic disks become simple volumes on the dynamic disks, and the other volumes become dynamic volumes of the corresponding type:
For detailed instructions for converting a basic disk to a dynamic disk, see the section "Upgrading a Basic Disk to Dynamic" in the previous chapter, "Disk Management."
There are particular considerations regarding dynamic disks and volumes on NetWare, Windows Server 2003, and Linux. See "Dynamic Disk and Volume Support on NetWare, Windows Server 2003, and Linux" for more information.
Dynamic volumes are volumes created on dynamic disks with Array Manager. On a dynamic disk, storage is divided into volumes instead of partitions. A dynamic volume consists of a portion or portions of one or more physical disks of any possible dynamic volume type. The five dynamic volume types are: simple, spanned, mirrored, striped, and RAID-5.
|
Note You cannot create a dynamic mirrored or RAID-5 volume if you are using a Windows NT Workstation or a Windows 2000 Professional machine. Also, dynamic volumes cannot be created on Microsoft Cluster Server (MSCS) shared resources (that is, cluster-enabled array disks that are attached to a PERC 3/DC controller). |
You can create any number of dynamic volumes in the unallocated space on a disk or create volumes that span two or more disks. Each volume on a disk can have a different file system, such as the file allocation table (FAT or FAT32) file system or the Microsoft Windows NT file system (NTFS).
Remember that any changes you make to your disk are immediately available. You do not need to quit Array Manager to save them or reboot your computer to implement them.
Topics covered in this section include:
There are particular considerations regarding dynamic disks and volumes on NetWare, Windows Server 2003, and Linux. See "Dynamic Disk and Volume Support on NetWare, Windows Server 2003, and Linux" for more information.
This section covers the five dynamic volume types:
There are particular considerations regarding dynamic disks and volumes on NetWare, Windows Server 2003, and Linux. See "Dynamic Disk and Volume Support on NetWare, Windows Server 2003, and Linux" for more information.
A dynamic simple volume consists of a single contiguous region on a disk. Dynamic simple volumes can be extended or mirrored. You can extend a simple volume within the same disk or onto additional disks. If you extend a simple volume across noncontiguous areas on the same disk or across multiple disks, it becomes a spanned volume. Dynamic simple volumes by themselves are not fault tolerant, but you can add data redundancy to a simple volume by mirroring it to another disk.
A dynamic spanned volume consists of two or more subdisks (contiguous disk regions) on one or more disks. With a spanned volume, you can combine sections of unallocated space from multiple dynamic disks into one large volume. The areas of unallocated space used to create spanned volumes can be different sizes. Spanned volumes are organized sequentially—that is, Array Manager sequentially allocates space on each disk until that disk is full and then continues with the next disk until the volume size is reached.
Spanned volumes can be extended. After a volume is extended, no portion of it can be deleted without deleting the entire spanned volume. Spanned volumes by themselves are not fault tolerant, but you can add data redundancy to a spanned volume by mirroring it to another disk.
When you create a dynamic volume, simple and spanned dynamic volumes are combined together in an option called concatenated.
|
Note Spanning volumes across multiple disk controllers is not recommended. |
Striped volumes are created by combining areas of free space on two or more disks into one logical volume. Data is divided into blocks and spread in a fixed order among all the disks in the volume. With a striped volume, data is written to multiple disks, similar to spanned volumes. However, striping writes files across all disks so that data is added to all disks at the same rate.
Striped volumes offer the best performance of all the disk management strategies. However, striped volumes do not provide fault tolerance. If a disk in a striped volume fails, the data in the entire volume is lost. You cannot mirror or extend a striped volume.
A mirrored volume is a volume that duplicates your data on two physical disks. A mirror provides redundancy by simultaneously writing the same data onto two separate volumes that reside on different disks. If one of the disks fails, data continues to be written to and read from the unaffected disk. You can mirror only a simple or spanned dynamic volume. Those two volume types combined together are known as concatenated.
A RAID-5 volume is a fault-tolerant volume with data and parity striped alternately across three or more physical disks. If one of the physical disks fails, data that was on the failed disk is recreated from the remaining data and parity on the other disks. RAID-5 volumes are a good solution for data redundancy in a computer environment in which most activity consists of reading data.
You need a minimum of three disks for a RAID-5 volume. Array Manager can accommodate up to 32 disks in a RAID-5 volume.
For more information about RAID strategies, see the section "Choosing RAID Levels" in the "Storage Management Concepts" chapter.
This section on creating a dynamic volume has some useful background information, as well as the detailed steps for successfully carrying out the procedure.
There are particular considerations regarding dynamic disks and volumes on NetWare, Windows Server 2003, and Linux. See "Dynamic Disk and Volume Support on NetWare, Windows Server 2003, and Linux" for more information.
Refer to the previous section, "Dynamic Volume Types," for a detailed description of the five dynamic volume types—simple, spanned, striped, mirrored, and RAID-5.
It is important to understand that the total volume size you indicate for the volume in the Create Volume wizard is the usable size for storing the data. Depending on the layout type, the volume can take more space on the disk. The size shown in the wizard for simple, striped, and spanned volumes is the actual volume size on the disk. A RAID-5 volume requires additional space for parity information, and a mirrored volume is a multiple of the space taken by the original plex of the volume.
Use the following procedure to create a dynamic volume.
There are particular considerations regarding dynamic disks and volumes on NetWare, Windows Server 2003, and Linux. See "Dynamic Disk and Volume Support on NetWare, Windows Server 2003, and Linux" for more information.
You are asked to select whether to create a partition or a volume. Make sure that the Dynamic volume button is highlighted.
You are asked to select the volume layout and size of volume to create.
If you are creating the volume on a Windows 2000 system, you will see a Query Max Size button. This button is not available on NT 4.0 systems. The Query Max Size button works differently, depending on whether you had a disk selected when you chose the Create Volume command.
You can choose a different disk for the volume you are creating. Click Modify to display the Modify Disks dialog box. Click the disk you want to change, and then click the arrow to select a disk from the drop-down list box.
|
Note It is strongly recommended that you leave the allocation unit size as Default. Selecting any other value requires an advanced knowledge of Windows file systems. You should thoroughly review the Microsoft Windows documentation before selecting any value other than Default. |
The capability of extending a simple or spanned volume, which is also known as Online Volume Expansion (OLVE), lets you increase the size of existing simple or spanned volumes. This is independent of the type of storage used, but is most beneficial when the virtual disks are first increased in size.
|
Note It is recommended that you use only hardware redundant (such as RAID-1 and RAID-5) virtual disks for spanned volumes. |
If you are planning to use OLVE with SAN storage, see the section that follows, "Considerations for Using OLVE with SAN Storage," before completing the instructions in this section.
You can extend a volume only if:
You cannot extend a volume if any of the following are true:
You can extend simple and spanned volumes on dynamic disks onto a maximum of 32 dynamic disks. Once a volume is extended, it cannot be mirrored or striped. No portion of a spanned volume can be deleted without deleting the entire spanned volume.
|
Note The Extend Volume option is disabled if the volume cannot be extended. |
Ignore any system messages that you may see while the volume extends. You do not have to run any special utilities. Examples of these messages are as follows:
System Process - Corrupt File
(X) The file or directory is corrupted and unreadable.
Please run the Chkdsk utility.
When using OLVE with SAN storage, the following need to be considered:
It is recommended that you use the following procedure for extending a simple or spanned dynamic volume across a new virtual disk in a SAN:
|
Note Never reassign a SAN LUN that is part of a spanned dynamic volume to another server or Free List. Data loss will occur. Also, it is strongly recommended that you use redundant RAID levels such as RAID-1 or RAID-5 for a virtual disk that will be used as part of a dynamic volume. |
This section gives more information on mirrored volumes and has the following topics:
There are two methods for creating a mirror:
|
Note You cannot extend or mirror a dynamic mirrored volume. |
The Add Mirror option is disabled if the volume cannot be mirrored or if there is no unallocated space on another dynamic disk large enough to mirror the volume.
Right-click a volume and then click Add Mirror.
You can remove or break a mirror. You should understand the differences between these two procedures.
Removing a mirror from a volume "removes" or destroys the data from the selected mirror and leaves the other mirror intact. After you remove a mirror, the space on the disk used by the removed mirrored volume becomes unallocated space. The remaining (no longer mirrored) volume becomes a simple volume on the disk.
Breaking a mirror creates two simple volumes with individual drive letters. Each volume contains the data on the mirror at the time the mirror was broken. The data is no longer redundant, but it remains intact.
Right-click the volume from which you want to remove a mirror and click Remove Mirror. You can then follow the instructions on your screen.
|
|
Caution Once you remove a mirror from a mirrored volume, the removed part of the mirror becomes unallocated space and the volume is no longer fault tolerant. All of the data on the removed mirror is deleted. |
Note that the remaining piece of the mirrored volume also becomes a simple volume. It retains the same drive letter of the original mirrored volume.
If there is a problem with one of the disks on which a mirrored volume resides, make sure that the disk is properly connected, and then try to bring that disk back online by first using Rescan. If Rescan does not work, then use the Reactivate Disk command. In most cases, Reactivate Disk will bring the disk online and all of the volumes will be healthy.
However, if any of the mirrors or plexes of the mirrored volume are still not healthy, you may need to resynchronize the unhealthy volume with the Reactivate Volume command. Resynchronizing makes the data consistent on the plexes of a mirrored volume.
Right-click the volume to bring up the context menu, and then click Reactivate Volume.
|
|
Caution When reactivating a volume, be aware that the volume's data is restored, even if it is stale, corrupt, or out-of-date. See "Reactivate a Dynamic Volume" for more information on the consequences of reactivating a volume. |
The mirrored volume should be brought back to a healthy state unless there is a serious problem with the volume or one of its disks has actually failed. In that case, see the next section, "Repairing a Mirrored Volume."
See also "Rescan to Update Information," "Reactivate a Disk," and "Reactivate a Dynamic Volume" in the "Troubleshooting" chapter.
When a disk on which a mirrored volume resides fails, the volume displays a Failed Redundancy status. The disk's name is changed to Missing, and an icon (X) appears on the Missing icon. The status of the disk will be Offline.
To repair a dynamic mirrored volume, you must first use Remove Mirror to remove the part of the volume that is on the missing disk and then use Add Mirror to add a new mirror, using space from another available dynamic disk.
The difference between reactivating a mirrored volume and repairing it is that reactivating is resynchronizing the volume to bring it to a normal state in its existing location, while repairing involves removing the damaged section of the volume and recreating it in another location on a healthy dynamic disk.
This section's topics include:
If there is a problem with one of the disks on which a RAID-5 volume resides, make sure that the disk is properly connected, and then try to bring that disk back online by first using Rescan. If Rescan does not work, then use the Reactivate Disk command. In most cases, Reactivate Disk will bring the disk online and all of the volumes will be healthy.
However, if the volume is still not healthy, try to resynchronize the unhealthy volume with the Reactivate Volume command. Right-click the volume to bring up the context menu, and then click Reactivate Volume.
|
|
Caution When reactivating a volume, be aware that the volume's data is restored, even if it is stale, corrupt, or out-of-date. See "Reactivate a Dynamic Volume" for more information on the consequences of reactivating a volume. |
The RAID-5 volume should be brought back to a healthy state unless there is a serious problem with the volume or one of its disks has actually failed. In that case, see the next section, "Repairing a RAID-5 Volume."
See also "Rescan to Update Information," "Reactivate a Disk," and "Reactivate a Dynamic Volume" in the "Troubleshooting" chapter.
You can repair a RAID-5 volume if it is in a state of Failed Redundancy, and if there is unallocated space on another dynamic disk available. To avoid data loss, you should attempt to repair the volume as soon as possible.
If you have a NetWare or Windows Server 2003 system or are planning on installing Linux, review the following information: