Multiple concepts that are easy to confuse and ignore in Vmware vsphere knowledge

0. Cloud and VirtualizationBaidu has a lot to explain about cloud computing, but it is estimated that many people still don't understand it. I personally think virtualization is like rain. virtualization aims to better implement cloud computing, with rain (virtualization ), the cloud has become closer to people's lives. The virtual host that has been born for a long time is actually a kind of cloud. People put the web page on it, but it is unclear that it is placed on that server, the current thunder network disk, offline download, online play while downloading is actually a cloud service1. What does IaaS PaaS SaaS mean and what is the difference?All three provide services, which are written as follows: Infrastructure Service IaaS integrates infrastructure (such as servers and storage), which is separated into different rooms for enterprises to rent like hotels, A typical example is the Amazon EC2 platform as a service PaaS: A program development platform and an operating system platform. This allows developers to write programs over the network, and users can also run programs on the platform, windows software as a service (SaaS): a software application model that provides software services on the Internet. In a typical example, the difference between MS and Office is that IaaS is like a highway, while PaaS is a car on a highway, saaS is the logistics company that ships goods2. Private cloud, public cloud, and hybrid cloudA public cloud provides services to both the public and the customer through the internet. A private cloud can be a data center in an enterprise and provide services to the Enterprise.3. vmotion and HAVmotion is a planned migration (hot migration ). Simply put, virtual machines on ESXi/ESX servers are automatically, semi-automatically, or manually (in combination with DRS) migrated to other ESXi/ESX hosts in the Cluster, the whole process does not require downtime, and the system will not be interrupted (some people are used to calling it seamless migration ). HA stands for High Availibility (High availability ). The out-of-service downtime protection function is simply implemented after A ESXi/ESX host in the same Cluster experiences abnormal downtime (power failure, hard disk failure, and so on, the VM is automatically started on another ESXi/ESX host in the Cluster. This ensures that the overall downtime does not exceed 5 to 10 minutes and can be protected by unexpected downtime (VM restart is required ). HA solves unexpected server downtime. For example, if your esx host suddenly crashes, the vm on this esx host will be restarted on other esx hosts to ensure business continuity.
Vomotion addresses planned downtime. For example, if your machine needs to upgrade the memory, you can first migrate the vm on the esx Server, and then power off and upgrade the memory. But your business (vm) does not pause.4. HA and FTFT: Fault Tolerance (Fault Tolerance ). The out-of-the-box downtime protection feature is further enhanced by implementing the abnormal ESXi/ESX host's FT-protected Virtual Machine without restarting on the basis of HA, it can run normally on other ESXi/ESX hosts in the Cluster without interruption. It is often referred to as "unplanned 0 down" protection. (FT currently requires a single Virtual Disk in vCPU and EagerZeroed Thick mode ).5. thick and thinAssume that you have a virtual machine with a disk of GB. the disk format is Thick backup, that is, Thick. Currently, the data volume is 20 GB. Because you are Thick, the virtual machine directly occupies GB of data storage space when it is created, and the space occupied by the data volume is within GB. At this time, if you create a snapshot and there are 50 GB of data changes after the snapshot, whether new or modified, the snapshot will occupy about 50 GB of space, at this time, the space occupied by this virtual machine in data storage is 100 GB + 50 GB. Of course, the memory swap file of the virtual machine also needs to occupy the data storage space. Assume that your VM is switched to the Thin standby mode, that is, the Thin mode. When the VM starts to use 20 GB of data, even if the VM disk is 100 GB, it only occupies 20 GB of data storage, when you create a snapshot and the number of snapshots changes by 50 GB, this virtual machine occupies 20 GB + 50 GB of data storage space. The above is the disk space usage, so what is the situation after the snapshot is deleted: the first Thick mode submits the 50 GB Data generated after the snapshot to the original GB disk file when deleting the snapshot. A large amount of disk I/O is generated, after completion, the virtual machine occupies 100 GB of disk space. The second Thin mode also submits 50 GB of data to the original disk file, the original GB disk file actually occupies only 20 GB of space in data storage. After submission, the virtual machine occupies 70 GB of disk space. The above is about the relationship between VMware disks and snapshots.6. Clone, template, and snapshotCloning is an identical backup on the basis of the original Virtual Machine. After cloning, the original virtual machine can continue to boot and run. If you select to convert the original host to a template, the original virtual machine is directly converted to a template. You can only use the template for batch deployment instead of starting or shutting down the host. The snapshot is used to save the current time state, which is equivalent to the time machine. If the Virtual Machine crashes in the future, you can play through7. VCB andVCB (vmware consolidated lidated backup) is a basic backup suite of vsphere in the past, but is now replaced by updated components. It is now referred to as the vmware data recovery technology, there is no need to install a backup proxy, just like a driver-free hardware. It also has the deduplication function, but it is still very limited, it is better to use third-party backup software.8. vapp and resource poolVApp: a set of n-layer application servers, such as web servers, middleware servers, and backend database servers, to form a web application service system. As a vApp, you can allocate and manage resources in a centralized manner, such as power-on and power-off at the same time. If you create an ovf template, You can conveniently deploy multiple web application systems with the same three-tier architecture. Resource pool: a resource pool that centrally allocates cluster resources. It mainly manages and allocates resources for servers in the pool, in particular, it reserves certain resources and limits the excessive use of resources (limit), and there are no special restrictions and regulations on virtual machines in the resource pool.9. Semi-virtualization and full virtualizationSemi-virtualization is achieved through the pre-modified client operating system kernel sharing the underlying host hardware. The advantage is that the semi-virtualized virtual machine operating system kernel can directly manage the underlying hardware, the performance is better than full virtualization technology. The disadvantage is that the kernel of the client operating system needs to be modified in advance, and the deployment convenience and flexibility are not enough, the compatibility is not good. full virtualization fully simulates the underlying hardware details through a virtualization logic layer hypervisor directly between the client and the host. The advantage is that the operating system kernel of the client does not require special configuration and is easy to deploy, flexibility and good compatibility. The disadvantage is that the kernel of the client operating system cannot directly manage the underlying hardware. The kernel manages the underlying hardware through the hypervisor, which requires a conversion overhead, and the performance is slightly weaker than that of semi-virtualization.10. Host and bare metalHOST: The hypervisor of the virtualization software needs to be installed in other operating systems. Advantage: the device is mainly managed by other underlying operating systems, such as windows, with rich drivers and support for many types of hardware. Disadvantage: hardware resource utilization is not high. The stability of all virtual machines depends on the stability of the underlying operating system. Bare metal: hypervisor of virtualization software can be directly installed on bare metal or hardware. Advantages: The hypervisor of virtualization software can directly manage the underlying hardware with high hardware utilization. The stability of all virtual machines depends on the hypervisor of virtualization software, and the stability is higher. Disadvantages: hypervisor of virtualization software currently does not support a wide range of hardware types, with fewer drivers, and many hardware cannot be used. The host type includes vmware workstation vmware vcenter. Microsoft windows2008hyper-V windows2008r2hyper-V and virtual PC RHEL5XEN and RHEL6KVM. The bare metal Parallels Desktop mode of virtualBox Parallels in oracle includes vmware esxi. Citrix's XenServer. Microsoft Hyper-V server. Rev Server. Oracle VM. Partition zzo Container of Parallels11. Three vswitch configuration modes: VST VGT ESTVirtual Switch Tagging (VST), External Switch Tagging (EST), and Virtual Guest Tagging (VGT) 1) EST-External Switch Tagging allocates Virtual Machine VLANs by dividing the ports of the vswitch into different VLANs. advantage: similar to managing physical environments, you do not need to configure VLANs on the ESXi server. disadvantage: virtual machines bound to the same physical port can only belong to the same vlan physical switch Configuration: switchport mode accessswitchport access vlan xx vSwitch configuration: No Virtual Machine (vm) configuration: no VLAN flag packets are processed on the physical switch. The port connecting the network adapter of the ESX host to the physical switch. The port group connecting to the vswitch must have its own vlan id set to 0. You can use an external switch to mark a VLAN. This is similar to a physical network. VLAN configurations are generally transparent to individual physical servers. When a packet arrives at the switch port, the label is removed when the packet leaves the switch port to go to the server. 2) VST-Virtual Switch Tagging: Configure multiple port groups in a vswitch to correspond to multiple VLANs. The advantage of a physical Switch to start trunk is recommended by VMware, one physical port in ESXi supports multiple VLANs. Disadvantages: complicated configuration of physical switches: switchport trunk encap dot1q vSwitch Configuration: in ESXi 'configuration'-'networking', select the corresponding vSwitch and add multiple port groups to the attribute. The vlan id of each port group corresponds to a vlan Virtual Machine (vm) Configuration: all VLAN flag packets are processed on the vswitch before they are sent to the ESX host. The network adapter of the ESX host must be connected to the relay port of the physical switch. The port group connecting to the vswitch must have a specified vlan id. In this mode, you provide a port group for each vlan on the vswitch, and then place the virtual network card of the virtual machine into the port group instead of directly connecting to the vswitch. The virtual machine port group marks all output frames and deletes all incoming frames. It also ensures that frames on a VLAN do not leak to different VLANs. 3) VGT-Vitual Guest Tagging identifies data of different VLANs through virtual machines. The advantage of enabling trunk on a physical switch is: applicable to special circumstances, such as the disadvantages of a virtual machine on a linux router: the virtual machine must support 802.1q physical switch configuration: switchport trunk encap dot1q vSwitch configuration: Change the vlan id of the port group in the vSwitch to 4095 Virtual Machine (vm) Configuration: all VLAN labels of the 802.1q trunk driver are processed by virtual machines. A 802.1Q VLAN relay driver must be installed on the virtual machine. When a data frame passes through a vswitch, the VLAN mark is stored between the virtual machine network stack and the external switch. Set the physical switch port to the Trunk port. You can install the 802.1Q trunking driver on a VM. When the frame passes through or reaches the VM, the flag will be retained between the VM network stack and the external switch.12. VMFS VMDK RDMVMFS is a file system of vmware. VMDK is a virtual hard disk file of vmware. RDM is a Raw Device Mappings native Device mapped to VMDK mode, and the LUN is mounted to storage by ESXI, in addition, it is stored in Datastore. The LUN will be formatted as VMFS, and the virtual hard disk of VM will be stored in the LUN of Datastore that has become VMFS in the format of VMDK files, in RDM mode, the LUN is regarded as an independent hard disk, that is, a LUN on the storage device. The LUN can be in various file formats, such as NTFS, EXT3, EXT4, and FAT32, depends on the operating system of the LUN. VM can use bit by bit to write the hard disk directly, but does not need to translate through hypervisor 13. What is the difference between cloud terminal and diskless workstation? A. differences in principles: operating principles of cloud terminals: in a shared working mode, all software is installed on the server and runs on the server. The running results are displayed on the cloud terminal, the cloud terminal is only responsible for display, keyboard, and mouse input, and does not run software. When you log on to different terminals, the server opens up a space for them. Therefore, different terminals run the same software without conflict. The effect of software running depends on the server. Operating principle of a diskless workstation: When a software is running on a diskless station, you need to load the software from the server to the local memory and then run it in the local memory. The running effect depends on the configuration of the diskless station. The configuration of diskless sites is complex and is an obsolete technology. B. differences in installation methods and performance 1. The configuration of diskless servers is extremely complex. You need to configure the NIC for each diskless server, you also need to reconfigure the server; 2. A large image is required for the diskless site, which is difficult to install, and the terminal does not need an image, it is much easier to install. 3. Because the diskless station is mostly self-assembled, the quality is not reliable, and the stability is very poor, it will soon begin to enter the maintenance period. 4. diskless stations can only be used within the LAN, it cannot be started across network segments, and cannot be connected to the wide area network. 5. Once the system is upgraded, each diskless site must be upgraded. This article is from the "prelude to reversal" blog.