|| Rate this content|
HP ProLiant DL185 G5 Storage Server - Maintaining the Storage Server
This document provides an overview of some of the components that make up the storage structure of the HP ProLiant Storage Server.
Storage management elements
Storage is divided into four major divisions:
Each of these elements is composed of the previous level's elements.
Storage management example
Figure titled Storage management process example depicts many of the storage elements that one would find on a storage device. The following sections provide an overview of the storage elements.
Figure 1: Storage management process example
Physical storage elements
The lowest level of storage management occurs at the physical drive level. Minimally, choosing the best disk carving strategy includes the following policies:
See Figure titled Configuring arrays from physical drives . With an array controller installed in the system, the capacity of several physical drives (P1-P3) can be logically combined into one or more logical units (L1) called arrays. When this is done, the read/write heads of all the constituent physical drives are active simultaneously, dramatically reducing the overall time required for data transfer.
NOTE: Depending on the storage server model, array configuration may not be possible or necessary.
Figure 2: Configuring arrays from physical drives
Because the read/write heads are simultaneously active, the same amount of data is written to each drive during any given time interval. Each unit of data is termed a block. The blocks form a set of data stripes over all the hard drives in an array, as shown in Figure titled RAID 0 (data striping) (S1-S4) of data blocks (B1-B12) .
Figure 3: RAID 0 (data striping) (S1-S4) of data blocks (B1-B12)
For data in the array to be readable, the data block sequence within each stripe must be the same. This sequencing process is performed by the array controller, which sends the data blocks to the drive write heads in the correct order.
A natural consequence of the striping process is that each hard drive in a given array contains the same number of data blocks.
NOTE: If one hard drive has a larger capacity than other hard drives in the same array, the extra capacity is wasted because it cannot be used by the array.
Drive failure, although rare, is potentially catastrophic. For example, using simple striping as shown in Figure titled RAID 0 (data striping) (S1-S4) of data blocks (B1-B12) , failure of any hard drive leads to failure of all logical drives in the same array, and hence to data loss.
To protect against data loss from hard drive failure, storage servers should be configured with fault tolerance. HP recommends adhering to RAID 5 configurations.
The table below summarizes the important features of the different kinds of RAID supported by the Smart Array controllers. The decision chart in the following table can help determine which option is best for different situations.
Further protection against data loss can be achieved by assigning an online spare (or hot spare) to any configuration except RAID 0. This hard drive contains no data and is contained within the same storage subsystem as the other drives in the array. When a hard drive in the array fails, the controller can then automatically rebuild information that was originally on the failed drive onto the online spare. This quickly restores the system to full RAID level fault tolerance protection. However, unless RAID Advanced Data Guarding (ADG) is being used, which can support two drive failures in an array, in the unlikely event that a third drive in the array should fail while data is being rewritten to the spare, the logical drive still fails.
Logical storage elements
Logical storage elements consist of those components that translate the physical storage elements to file system elements. The storage server uses the Window Disk Management utility to manage the various types of disks presented to the file system. There are two types of LUN presentation: basic disk and dynamic disk. Each of these types of disk has special features that enable different types of management.
Logical drives (LUNs)
While an array is a physical grouping of hard drives, a logical drive consists of components that translate physical storage elements into file system elements.
It is important to note that a LUN may span all physical drives within a storage controller subsystem, but cannot span multiple storage controller subsystems.
Figure 4: Two arrays (A1, A2) and five logical drives (L1 through L5) spread over five physical drives
NOTE: This type of configuration may not apply to all storage servers and serves only as an example.
Through the use of basic disks, you can create primary partitions or extended partitions. Partitions can only encompass one LUN. Through the use of dynamic disks, you can create volumes that span multiple LUNs. You can use the Windows Disk Management utility to convert disks to dynamic and back to basic and to manage the volumes residing on dynamic disks. Other options include the ability to delete, extend, mirror, and repair these elements.
Partitions exist as either primary partitions or extended partitions and can be composed of only one basic disk no larger than 2 terabytes (TB). Basic disks can also only contain up to four primary partitions, or three primary partitions and one extended partition. In addition, the partitions on them cannot be extended beyond the limits of a single LUN. Extended partitions allow the user to create multiple logical drives. These partitions or logical disks can be assigned drive letters or be used as mount points on existing disks. If mount points are used, it should be noted that Services for UNIX (SFU) does not support mount points at this time. The use of mount points in conjunction with NFS shares is not supported.
When planning dynamic disks and volumes, there is a limit to the amount of growth a single volume can undergo. Volumes are limited in size and can have no more than 32 separate LUNs, with each LUN not exceeding 2 terabytes (TB), and volumes totaling no more than 64 TB of disk space.
The RAID level of the LUNs included in a volume must be considered. All of the units that make up a volume should have the same high-availability characteristics. In other words, the units should all be of the same RAID level. For example, it would not be a good practice to include both a RAID 1+0 and a RAID 5 array in the same volume set. By keeping all the units the same, the entire volume retains the same performance and high-availability characteristics, making managing and maintaining the volume much easier. If a dynamic disk goes offline, the entire volume dependent on the one or more dynamic disks is unavailable. There could be a potential for data loss depending on the nature of the failed LUN.
Volumes are created out of the dynamic disks, and can be expanded on the fly to extend over multiple dynamic disks if they are spanned volumes. However, after a type of volume is selected, it cannot be altered. For example, a spanning volume cannot be altered to a mirrored volume without deleting and recreating the volume, unless it is a simple volume. Simple volumes can be mirrored or converted to spanned volumes. Fault-tolerant disks cannot be extended. Therefore, selection of the volume type is important. The same performance characteristics on numbers of reads and writes apply when using fault-tolerant configurations, as is the case with controller-based RAID. These volumes can also be assigned drive letters or be mounted as mount points off existing drive letters.
The administrator should carefully consider how the volumes will be carved up and what groups or applications will be using them. For example, putting several storage-intensive applications or groups into the same dynamic disk set would not be efficient. These applications or groups would be better served by being divided up into separate dynamic disks, which could then grow as their space requirements increased, within the allowable growth limits.
NOTE: Dynamic disks cannot be used for clustering configurations because Microsoft Cluster only supports basic disks.
File system elements
File system elements are composed of the folders and subfolders that are created under each logical storage element (partitions, logical disks, and volumes). Folders are used to further subdivide the available file system, providing another level of granularity for management of the information space. Each of these folders can contain separate permissions and share names that can be used for network access. Folders can be created for individual users, groups, projects, and so on.
File sharing elements
The storage server supports several file sharing protocols, including Distributed File System (DFS), Network File System (NFS), File Transfer Protocol (FTP), Hypertext Transfer Protocol (HTTP), and Microsoft Server Message Block (SMB). On each folder or logical storage element, different file sharing protocols can be enabled using specific network names for access across a network to a variety of clients. Permissions can then be granted to those shares based on users or groups of users in each of the file sharing protocols.
Volume Shadow Copy Service overview
The Volume Shadow Copy Service (VSS) provides an infrastructure for creating point-in-time snapshots (shadow copies) of volumes. VSS supports 64 shadow copies per volume.
Shadow Copies of Shared Folders resides within this infrastructure, and helps alleviate data loss by creating shadow copies of files or folders that are stored on network file shares at pre-determined time intervals. In essence, a shadow copy is a previous version of the file or folder at a specific point in time.
By using shadow copies, a storage server can maintain a set of previous versions of all files on the selected volumes. End users access the file or folder by using a separate client add-on program, which enables them to view the file in Windows Explorer.
Shadow copies should not replace the current backup, archive, or business recovery system, but they can help to simplify restore procedures. For example, shadow copies cannot protect against data loss due to media failures; however, recovering data from shadow copies can reduce the number of times needed to restore data from tape.
Using storage elements
The last step in creating the element is determining its drive letter or mount point and formatting the element. Each element created can exist as a drive letter, assuming one is available, and/or as mount points on an existing folder or drive letter. Either method is supported. However, mount points cannot be used for shares that will be shared using Microsoft Services for Unix. They can be set up with both but the use of the mount point in conjunction with NFS shares causes instability with the NFS shares.
Formats consist of NTFS, FAT32, and FAT. All three types can be used on the storage server. However, VSS can only use volumes that are NTFS formatted. Also, quota management is possible only on NTFS.
Clustered server elements
Select storage servers support clustering. The HP ProLiant Storage Server supports several file sharing protocols including DFS, NFS, FTP, HTTP, and Microsoft SMB. Only NFS, FTP, and Microsoft SMB are cluster-aware protocols. HTTP can be installed on each node but the protocols cannot be set up through cluster administrator, and they will not fail over during a node failure.
CAUTION: AppleTalk shares should not be created on clustered resources as this is not supported by Microsoft Clustering, and data loss may occur.
Network names and IP address resources for the clustered file share resource can also be established for access across a network to a variety of clients. Permissions can then be granted to those shares based on users or groups of users in each of the file sharing protocols.
Network adapter teaming
Network adapter teaming is software-based technology used to increase a server's network availability and performance. Teaming enables the logical grouping of physical adapters in the same server (regardless of whether they are embedded devices or Peripheral Component Interconnect (PCI) adapters) into a virtual adapter. This virtual adapter is seen by the network and server-resident network-aware applications as a single network connection.
HP Systems Insight Manager
HP SIM is a web-based application that allows system administrators to accomplish normal administrative tasks from any remote location, using a web browser. HP SIM provides device management capabilities that consolidate and integrate management data from HP and third-party devices.
NOTE: You must install and use HP SIM to benefit from the Pre-Failure Warranty for processors, SAS and SCSI hard drives, and memory modules.
For additional information, refer to the Management CD in the HP ProLiant Essentials Foundation Pack or the HP SIM website ( http://www.hp.com/go/hpsim ).
Management Agents provide the information to enable fault, performance, and configuration management. The agents allow easy manageability of the server through HP SIM software, and thirdparty SNMP management platforms. Management Agents are installed with every SmartStart assisted installation or can be installed through the HP PSP. The Systems Management homepage provides status and direct access to in-depth subsystem information by accessing data reported through the Management Agents. For additional information, refer to the Management CD in the HP ProLiant Essentials Foundation Pack or the HP website ( http://www.hp.com/servers/manage ).
NOTE: The HP Systems Management Homepage is not supported on the HP ProLiant 100-series servers.