1.LVM (Logical Volume Manager): Logic scrollbar administrator, LVM can partions several entities (or disk) through a combination of software seems to become an independent large disks (VG), and then this big disk after segmentation can be used again become separated groove (LV), it is possible to mount the final use.
2.PV (Physical Volume): entity scrollbar, we need to adjust the actual partition system identifier (system ID) called 8e (LVM identifier), and then through the pvcreate command will turn it into the bottom of the LVM entity scrollbar (PV), before it could take advantage of these PV. Adjustment system ID method is fdisk.
3.VG (Volume Group): scrollbars group, LVM disk is to integrate a number of large PV to VG, VG is a combination of large capacity LVM disk, the disk is determined by the large PE, because each can contain only up to VG 55534 PE. If you are using LVM preset parameters, a VG up to 25G capacity
4.PE (Physical Extend): entity extension block, LVM default to 4MB of PE blocks, and the LVM VG can only contain a maximum of 65 534 PE, so the default LVM VG will be 4 * 65534/1024 = 256G .PE LVM is the smallest storage blocks, as it is like the file system block size.
5.LV (Logical Volume): Logic scrollbar, VG will eventually be cut to LV, the LV is the last thing that can be formatted similarly split groove used. To facilitate use LVM to manage their systems, so the LV device file name is usually designated as / dev / vgname / vname styles.
6.LVM elastically changed filesystem capacity, which is performed by exchanging data conversion PE, the PE within the original LV transferred to another device to reduce the capacity of the LV, or other PE was added to increase the capacity of the LV . Relations VG, LV, PE's
PE VG within'll give LV yellow line portion, if the future of the VG want to expand it, plus other PV can. The LV If you want to expand, then, is not used by PE in the VG to the expansion.
7. Through PV, VG, after LV planning, then you can use mkfs to format your LV can take advantage of file system, the file system capacity in the future can also be expanded or reduced, and there is not data It will be affected.
8. When data is written to LV, there are two ways to write hard disk:
If the time / devhda1, / dev / hdb1 partition these two were added to the VG, and the entire VG only a LV, so-called linear model is:: 1) linear mode when the / dev / hda1 capacity runs out / dev / hdb1 it will be used to, which is the recommended mode
2) interleaved mode: interlace mode is to split the data into a sum of two parts, written to / dev / hda1, / dev / hdb1 mean, a data written to two disks, in theory, performance will be better.
9. Basically, LVM's main use is in the realization of an elastic adjustment capacity of the file system, rather than to establish a performance-based disk, so we should take advantage of the LVM can use flexible management of the entire partition size, rather than focusing on performance. Therefore, LVM is the default read-write mode linear mode, if you use interlaced mode, should be noted that, when any partition is damaged, all data will be corrupted. If you want to emphasize efficiency and backup with RAID or better.
10.LVM System Snapshot: Snapshot is speaking at the time of system information recorded, just like photographic records in general, if there are any future changes to the data, the original data to the snapshot area to avoid moving, what-have altered the region by the snapshot file system sharing.
Originally build snapshot of the system status, LVM will reserve an area (on the left side three PE data graph, so you'll see the contents of the file system snapshot area is exactly the same after a while until the system works, it is assumed a data area is changed, the changes before the system moves the data area to the snapshot area, so taking up a PE called a, B and others to block the I or shared file systems.