This week we learned about the installation of Windows Server 2012. It was a pretty wide-ranging lesson, from how, when and why to implement the three types of user interface from a full Windows experience with Internet Explorer included; a GUI of server management tools; or command line access only.
Control of a server using Windows Power Shell commands from a command line only obviates the need for a graphical interface using up precious clock cycles and closes potential server vulnerabilities by avoiding the use of a graphical interface. This type of installation is knows as Server Core.
We also learned a lot about the modern tools available in the re-thought Windows Server 2012. NIC teaming allows an administrator to gang each interface together to either aggregate performance or provide failover capabilities for each interface.
The training course also covered the usage of storage devices and how to connect and install them on the server. While they went into great depth to cover the ability of the operating system to provide the traditional types of software RAID arrays of mirroring, striping and parity to provide different types of volumes. These can be easily understood:
A simple volume is a single and contiguous block of space of a single hard drive. There is no failover tolerance in this mode, i.e., any hard drive dying loses all data on it.
A spanned volume combines areas from two or more disks into one unit. This has the advantages of combining scraps of space the might otherwise go unused, but like the simple volume provides no protection against hardware failure.
A striped volume breaks data into units and then stores those units across various disk drives that are available to the system. This can be set up on two or more disks as long as the size of the partition on each disk is the same. This provides an increase in performance but no fault toleration. The acronym RAID 0 applies to this type of volume.
A striped volume with parity acts much like the striped volume above, but adds parity information to each drive used. This has the benefit that data can be recovered by the combination of the existing data with the existing parity information surviving on the non-failed devices to recreate the dataset across all devices in the array. This type of array is analogous to RAID 5.
The mirrored volume stores two copies of each file on two different disk drives. This has the effect of halving storage capacity, but provides fault toleration in case of hardware failure. This type of array is also known as RAID 1.
Having gone into some length about this, the instructor then dropped the bombshell that a sysadmin most likely would rarely if ever use this software RAID implementation to manage hard drives. A new tool, Storage Pools, helps the system administrator to aggregate available storage devices or at the same time provide failover services. This more closely mimics traditional hardware RAID services, but is more flexible and does not usually require the admin to configure each device separately.
Likewise, server aggregation or failover tolerance is provided by the tool Server Pool. This allows the admin to pool servers in groups, and then administer the group as if they were working on an individual machine. Server groups, or Pools, can then be tasked with server roles and features.
A tool that interested me extremely is Windows RM, or Remote Manager. Windows RM allows an administrator to push commands in real time to another server (or any joined machine) on the workgroup or domain. The advantage of this is that any command that can be given from the command line can be immediately sent to the remote system without requiring a reboot.
There are many more features that I am not able to cover in this brief synopsis such as virtual hard disks and Server 2012's ability to import and export disks and servers.
This week was an overwhelming but exhilarating introduction to Windows Server 2012. I'm looking forward to next week's lesson!
No comments:
Post a Comment