A block storage device is the physical, logical, or virtual storage media available to a server. A device can be directly attached to the server or connected via storage networking protocols such as Fibre Channel and iSCSI.
For OES 11 SP3, NSS recognizes device sizes up to 2E64 sectors (that is, up to 8388608 petabytes (PB) based on the 512-byte sector size).
NSS allows you to initialize devices to use either the DOS partition table format or the GPT (GUID Partition Table) format for a given device. When you create pools, you can combine space from devices that are formatted using any method.
The DOS partition table scheme supports devices up to 2TB in size. It allows up to four partitions on a device.
The GPT partition table scheme supports device sizes up to 2E64 sectors (that is, up to 8388608 petabytes (PB) based on the 512-byte sector size). It allows up to 128 partitions per disk. Each of its disks partitions is a logical device that is identified by a unique 128-bit (16-byte) GUID.
IMPORTANT:Pools created on GPT initialized devices are not backward compatible with versions earlier to OES 11.
Different manufacturers report device sizes differently. The actual device size varies with the hardware design and the applications and software drivers that manage the device. Many vendors report sizes using a definition where 1 TB = 10E12 bytes = 1,000,000,000,000 bytes. Space can also be consumed by metadata that is added to manage the device. The location on the device where the metadata is stored can also vary by hardware manufacturer and software vendor. After you format the drive, yet another size might be reported. Third-party product documentation might state the maximum size limits of devices it supports before or after making accommodations for any management data or space lost to formatting. The size of devices you ultimately carve out for use with NSS depends on all these factors.
IMPORTANT:Make sure to refer to the documentation of the device manufacturer, application vendor, and software driver vendor for other limitations on the device size.
The following are examples of common types of devices:
Server disks include physical disks on the server or logical disks carved from the server disk.
Physical or logical disks can be directly attached to the server as individual devices or in a storage array.
A LUN (logical unit number) can be either a physical or a logic disk drive. Refer to the iSCSI SAN or Fibre Channel SAN documentation for information about creating and managing LUNs for your SAN implementation.
A metaLUN is a controller-managed group of multiple LUNs or of multiple hardware RAIDs that are striped or concatenated together to be presented as a single LUN device to the server. Refer to the hardware manufacturer’s documentation for information about creating metaLUNs.
An iSCSI device is a remote target disk or tape drive on an iSCSI disk server that is made available across an IP network by iSCSI initiator software running on the server. After connecting to the disk server, you can view the devices in the Devices list and add NSS pools and volumes as you would with any device.
For information about managing and using iSCSI devices on OES, see Mass Storage over IP Networks—iSCSI
in the SLES 11 SP4 Storage Administration Guide. See also the Linux iSCSI Project.
A RAID (redundant array of independent disks) is a logical device that combines space from multiple devices by using special hardware, software, or both. Data is striped or replicated across all member devices to improve data reliability, increase I/O performance, or provide device fault tolerance. All RAID types require configuration using a RAID management tool made for the specific hardware or software used in the RAID.
In a hardware RAID, the RAID functionality and management are in firmware within the storage cabinet. Refer to the hardware manufacturer’s documentation for information about creating hardware RAIDs.
Controller RAID devices are also known as BIOS RAIDs, fakeRAIDs, hostRAID, and quasi-hardware RAIDs.
In controller RAIDs, the functionality and management are in the HBA or controller BIOS/firmware. If the controller does not contain an on-board CPU resource to use for RAID management, the controller RAID consumes server CPU resources to manage the RAID.
Refer to the hardware manufacturer’s documentation for information about configuring Controller RAID devices. For information about using Controller RAIDs with OES, see TID 3626577: BIOS RAID Support in the Novell Support Knowledgebase.
Software RAIDs are controlled by special software in the server's OS such as in the HBA driver or in upper level module such as NSS. Software RAIDs consume CPU resources to manage the RAID.
For information about creating and managing NSS software RAIDs, see Section 13.0, Managing NSS Software RAID Devices.
You can optionally use Linux tools to create and manage Linux software RAIDs. Linux software RAIDs must be initialized with either a DOS or GPT partition format in order to be used by NSS.
If there are multiple connection paths between a device’s hardware controller and the server, each path presents a given device to the server as a separate device. You must use a multipath management tool to resolve the multiple apparent devices to a single multipath device. Use the multipath device UUID or alias when you are creating NSS pools and volumes. Multipath tools also provide automatic path management for path failover, failback, and reconfiguration.
Use the Linux multipath I/O tools to create the multipath device.
Removable media devices include CDs, DVDs, or CD/DVD image files. Removable media are mounted as Linux POSIX file systems. Use Linux native tools to manage removable media.
In a Xen virtual environment, you use the Virtual Machine Manager in YaST to allocate storage devices from the host to the virtual machine. The devices that you want to use for the NSS file system on the guest machine must be less than 2 TB (where 1 TB = 2E40 bytes) if the guest operating is OES 2 SP3 or earlier versions. Devices of 2 TB or larger are supported from OES 11 onwards. For information about storage considerations in virtual environment, see Section 6.0, Using NSS in a Virtualization Environment.
The following table describes the type of information available for each device by viewing the Device Details page.
Table 10-1 Explanation of Device Details
Device Detail |
Description |
---|---|
Name |
The device name assigned by the device manager. |
Major Number Minor Number |
The device identity in major:minor format. Major and minor numbers are associated with the device special files in the /dev directory and are used by the operating system to determine the actual driver and device to be accessed by the user-level request for the special device file. |
Partitioning Type |
DOS - The device has the DOS partitioning scheme. GPT - The device has the GPT partitioning scheme. LVM2 volume - The device has no partition table because there is an LVM volume at the beginning of the disk. This is the case when we create a Clustered Linux Volume. CSM - Unknown - The device is not recognized by NSS. |
Shareable for Clustering |
The attribute of a device that indicates whether the selected device can be shared by multiple computers in a cluster solution. |
Capacity |
The total available storage space of the selected device. |
Used Space |
The amount of space on the device that is currently in use by partitions, including Novell partitions for NSS as well as native Linux partitions. |
Free Space |
The total amount of space on the device that is currently not in use. |
Pools |
The drop-down list shows all pools that exist on this device. To view a pool’s details or to manage a pool, select the pool from the list, then click the View Details icon to go to the Pools page for that pool. |
Number of Pools |
The total number of pools that use this device. |
Partitions |
The drop-down list shows all partitions that exist on this device. To view a partition’s details, select the partition from the list, then click the View Details icon to go to the Partition Information page for that partition. |
Mirror Status |
For a RAID1 device, this field shows its status:
|