2.1 Install and configure server operating systems
- Determine server role/purpose – the purpose of each installed server should be determined before its own purchase. Each server can have its own specific usage, which answer to specific business need. You can have a web server or database server role – each with its own requirements regarding resources – number of CPUs, disk capacity, speed, capacity of memory etc.
- Update firmware – each server hardware component has its own program stored in some kind of non volatile memory on the component itself, that tells the component how to behave and provides basic very low level operation. The firmware like any other software needs to be maintained – in order to fix bugs or present new features. In server environment it is especially crucial as some of the bugs can impact many end users.
- BIOS/UEFI configuration – BIOS and UEFI are described in Server Architecture. BIOS is the very first system that communicates with the HW components talking with its firmwares. UEFI is the successor to BIOS and adding some new features -ability to use disks with more than 2TB with GPT (boot doesn’t depend on boot sector), advanced pre-OS environment (often with network capability), 64bit drivers etc.
- Boot order – from the perspective of Server Administration, boot order is an information inside Setup of BIOS that tells BIOS how to proceed with boot from various devices – in what order the BIOS/UEFI should search the devices for boot information. For example the first can be CD/DVD drive, the second flash disk, the third HDD etc. – different situation requires different boot order or you can select specific device from the boot menu.
- Disk preparation – before an Operating system can be installed anywhere, you should prepare the drive.
- RAID setup – RAID should be set up first, before anything else. Usually this can be done through RAID controller’s firmware, which allows you to enter some kind of management utility. The management utility then allows you to create a RAID out of available disks. (for RAID description please refer to Storage section of Server+). Proper RAID setup will ensure that drives are combined in a RAID of some number and presented as one volume by the RAID controller.
- Partitioning – partitioning is next and can be done by the Operating System installer medium – basically it means you can split the volume (it can be one disk or it can be some RAID configuration therefore I refer here to it as a volume) to some logical parts. You can partition the whole volume as one big partition or you can split it to some separated parts. It allows you then to have multiple file systems – each file system on separate partition and for example you can have more Operating Systems on one volume etc. (not very usable in case of servers). Other usage is to create one partition for operating system and second for data, so you do not mix everything on one partition. In the world of servers you will mostly have in mind this when you are selecting what kind of disks you already have in the server and won’t partition one volume for everything. Mostly you will have more volumes – for example RAID 1 volume out of 2 disks for OS (some kind of best practice) and then some kind of performance/fault-tolerant volume out of the rest of the disks, where you will store your application data and outputs.
- Formatting – refers to putting or creating a file system of some type on the partition. File system tells the OS how the data is store and how it can be retrieved.
- File system type – there are many types of file systems. Usually they are more or less tied to some particular Operating system (NTFS – Windows, ext4 – Linux etc.)
- Ext 2, 3, 4 – extended file system for Linux. Version 2 didn’t have a journal system, version 3 has a journal system and in ext4 you can switch the journaling off. Journaling is a method that most of current and advanced file systems use – it allows to write an information about a data operation before the data operation happens. This adds to resilience of the data in case the OS crashes.
- NTFS – New Technology File System – Windows OS file system. It has many advantages over older Microsoft file systems (FAT16/32 etc.) like journaling, advanced data structures, improved security with built-in ACLs etc.
- FAT32
- ReiserFS
- UFS
- VMFS
- ZFS
- Limits of file systems:
-
File system name Maximum file size Maximum volume size Ext2 16GB – 2TB 2TB – 32TB Ext3 16GB – 2TB 2TB – 32TB Ext4 16TB – 16TB 1EB NTFS 16TB 256TB FAT32 4GB minus 1 byte 32GB ReiserFS 8TB (newer version) / 4GB (older version) 16TB UFS 8ZB 8ZB VMFS (5) 62TB 64TB ZFS 16EB 256ZB
- Swap
- Configure host name
- Local account setup
- Connect tnetwork
- Join domain/directory
- Address security concerns
- Patching
- OS hardening
- Compliance tcompany procedures/standards
- Enable services
- Install features/roles/applications/drivers
- Performance baseline
- Server optimization
- Swap or pagefile optimization
- Unattended/remote installations
- Deploying images and cloning
- Scripted installs
- PXE boot
- TFTP
2.2 Compare and contrast server roles and requirements for each
- Web server
- Application server
- Directory server
- Database server
- File server
- Print server
- Messaging server
- Mail server
- Routing and remote access server
- Network services server
- DHCP
- DNS/WINS
- NTP
2.3 Given a scenario, use access and control methods tadminister a server
- Local hardware administration
- KVM
- Serial
- Virtual Administration console
- Network-based hardware administration
- KVM over IP
- IL
- iDRAC
- Network-based operating system administration
- RDP
- SSH
- VNC
- Command line / shell
2.4 Given a scenario, perform proper server maintenance techniques
- Change management
- Patch management
- Operating System updates
- Application updates
- Security software updates
- Firmware updates
- Device drivers updates
- Compatibility lists
- Operating systems
- Hardware
- Applications
- Testing and validation
- Outages & Service Level Agreements
- Scheduled downtime
- Unscheduled downtime
- Impact analysis
- Client notification
- MTTR
- Performance monitoring
- CPU utilization
- Memory utilization
- Network utilization
- Disk utilization
- Disk IOPS
- Storage capacity
- Comparison against performance baseline
- Processes and services monitoring
- Log monitoring
- Hardware maintenance
- Check system health indicators
- LEDs
- Error codes
- Beep codes
- LCD messages
- Replace failed components
- Fans
- Hard drives
- RAM
- Backplanes
- Batteries
- Preventative maintenance
- Clearing dust
- Check proper air flow
- Proper shut down procedures
- Check system health indicators
- Fault tolerance and high availability techniques
- Clustering
- Active/active
- Active/passive
- Load balancing
- Round robin
- Heartbeat
- Clustering
2.5 Explain the importance of asset management and documentation
- Asset management
- Licensing
- Labeling
- Warranty
- Life cycle management
- Procurement
- Usage
- End of life
- Disposal/recycling
- Inventory
- Make
- Model
- Serial number
- Asset tag
- Documentation
- Service manuals
- Network diagrams
- Architecture diagrams
- Dataflow diagrams
- Recovery documentation
- Baseline documentation
- Change management policies
- Service Level Agreement
- Server configuration
- Secure storage of sensitive documentation
2.6 Explain the purpose and operation of virtualization components
- Hosts and guests – hosts machines are usually servers that run a hypervisor software, which allows guest(s) run on top of it. This server allows to host multiple guests with different operating system. The guests can be also named VMs – Virtual Machines.
- Management interface for virtual machines
- Hypervisor – is the software that allows to host multiple guest operating system.
- Type I – bare metal hypervisors. These hypervisors are installed directly on the server disks and have their own management interfaces to create and support guests. Examples are Citrix XEN, VMware ESXi or Microsoft Hyper-V.
- Type II – hosted hypervisor. These hypervisors run above an existing Operating System. For example you have a desktop PC with Windows 10 OS and you switch-on Hyper-V feature which allows you to create more OS running on top of your Hyper-V application in your current operating system. In this case hypervisor is not run directly from the disks but rather is loaded from the Operating System after the Operating system starts.
- Hybrid – Not sure what CompTia wants to hear here. Google isn’t very clear about answers to question of hybrid hypervisor – maybe they refer to a hybrid cloud – combination of on-premise installed hypervisors that can connect to private clouds and cooperate with them or maybe they refer to a container style virtualization like Docker.
- Hardware compatibility list – most vendors will say if their HW is capable of running virtualized environment or not.
- BIOS/UEFI compatibility and support – this feature has to be enabled for most of the bare metal hypervisors – one would enable this feature and the hypervisor is able to support the guests with HW resources and various specific features.
- CPU compatibility support – one of the most crucial feature if not the most. The CPU should be compatible with running virtualized environments. You can google for this information if you try to look for the model number of your processor – for example this Intel i5 6600K – https://ark.intel.com/products/88191/Intel-Core-i5-6600K-Processor-6M-Cache-up-to-3_90-GHz.
- AMD-V / Intel VT – these are names of hardware assisted virtualization support for the biggest processor vendors. Usually one has to enable this feature in the BIOS/UEFI in order to make it work (and hypervisors won’t run if this feature is disabled).
- Resource allocation between Guest and Host – each host a specific number of resources that can be allocated to guests (virtual machines). Resources in this case can be allocated either statically or dynamically based on different situation. Static allocation means that each guest has a maximum number of resources set that it can use. Dynamic allocation works better in larger numbers – for example VMware has a DRS feature where a hosts resources become part of the cluster resources – the cluster then decides where the VM should be located so it runs smoothly resources wise and resources themselves are based on shares configured for each virtual machine.
- CPU – regarding CPU – there is a lot of that can be set – number of sockets, number of cores, CPU affinity etc. What is interesting is how the hypervisor handles the distribution and what does it mean in terms of difference between set of a socket and of a core. The socket vs core setting – assuming for example we have a 2 socket host with both sockets used by a processors of total 16 cores. We can add to a VM up to 16 cores (not assuming HT) – if I add 1 socket with 2 cores, VM uses 2 cores; if I add 2 socket each with 1 core, VM uses the same – 2 cores. Obviously it seems there is no need to create virtual sockets with virtual cores. The catch is in the OS that will be installed on the VM – what it can see/use. Some OSes (like Windows Server 2008 R2) can lock the maximum amount of sockets and cores it can use based on the version / license (Standard/Enterprise/Datacenter) – for example for the Standard there is a limitation to 8 sockets. So if you would like the VM to use 16 cores of the host and would set up 16 sockets with 1 core each, you would end up with only 8 displayed and usable CPUs in the OS. But if you would set up 8 sockets, 2 cores each you would see full 16 CPUs inside the OS.
- Storage – regarding storage there are 2 main attributes that can be set regarding resources – capacity and IOPS (speed). Capacity is like in any regular disk – it determines the number of data that can be stored. However in case of virtual environment it tends to be a easier when it comes to increase of disk space – every disk is stored as a file, so increasing the size of the file can be just a few clicks. Regarding speed – the Hypervisor can limit the number of IOPS the VM can produce towards the datastore it resides on.
- Memory – memory is quite straightforward – usually there is some limit and possibility to reserve the amount otherwise it depends on the hypervisor to allocate additional memory in case of a need (and so the memory can be depleted for other VMs, if there is a high need from one VM).
- Network connectivity
- Direct Access (Bridging) vs NAT
- Direct Access (bridgining) enables the VM to use NIC of the host directly – be in the same network as a host (has IP address from network of the host)
- NAT – host NIC will serve as an interface for NAT and VMs will be located in their separate network. Packets coming from the VMs network will be translated to the IP address of the Host.
- Direct Access (bridgining) enables the VM to use NIC of the host directly – be in the same network as a host (has IP address from network of the host)
- Virtual NICs – is a representation of a network interface card in a VM. VM can have a s a physical server multiple vNICs for the purpose of fault tolerance or load balancing.
- Virtual switches – is a switch created inside the hypervisor and providing switching functionalities. It is dependent on the hypervisor’s vendor – VMware has Standard and Distributed virtual switches, Hyper-V has Internal (NAT), External (direct access = bridged) and Private (separated switch – only for VMs communication with each other) switches.
- Direct Access (Bridging) vs NAT
- Video – hypervisor doesn’t have to necessarily serve only for server type virtual machines. It can be used also for virtualizing desktops (aka VDI). For this purpose you have to have some kind of possibility of adding a video card to your VMs that will serve for all kind of graphical demands coming from virtualized dekstops (flash videos etc.). Adding a vGPU will offload a ton from your processors.