Bạn đang đọc: Disk mirroring
Disk mirroring differs from charge shadowing that operates on the file tied, and phonograph record snapshots where data images are never re-synced with their origins .
Overview [edit ]
It is recognized that disks are an inherently unreliable component of computer systems. Mirroring is a proficiency to allow a system to automatically maintain multiple copies, or a double stand-in ( meaning that the datum is pleonastic on all arduous drives that exist in the mirror ) of data so that in the event of a phonograph record hardware failure a system can continue to process or cursorily recover data. mirror may be done locally where it is specifically to cater for harrow undependability, or it may be done remotely where it forms part of a more sophisticated catastrophe recovery scheme, or it may be done both locally and remotely, specially for high handiness systems. Normally data is mirrored onto physically identical drives, though the serve can be applied to legitimate drives where the implicit in physical format is hidden from the mirror action. typically, mirroring is provided in either hardware solutions such as phonograph record arrays, or in software within the operate system ( such as Linux mdadm and device plotter ). [ 1 ] [ 2 ] Additionally, charge systems like Btrfs or ZFS provide integrated data mirroring. [ 3 ] [ 4 ] There are extra benefits from Btrfs and ZFS, which maintain both data and metadata integrity checksums, making themselves able of detecting badly copies of blocks, and using mirrored data to pull up data from correct blocks. [ 5 ] There are several scenarios for what happens when a harrow fails. In a hot swap system, in the event of a disk failure, the system itself typically diagnoses a disk failure and signals a bankruptcy. sophisticated systems may mechanically activate a hot standby phonograph record and use the remaining active disk to copy live data onto this magnetic disk. alternatively, a newly disk is installed and the datum is copied to it. In less sophisticate systems, the system is operated on the remaining disk until a spare harrow can be installed.
Xem thêm: Rims vs. Wheels: What’s the Difference?
The copy of data from one side of a mirror pair to another is called rebuilding or, less normally, resilvering. [ 6 ] Mirroring can be performed site to web site either by rapid data links, for example fiber eye links, which over distances of 500 megabyte or thus can maintain adequate performance to support real-time mirror. Longer distances or slower links maintain mirrors using an asynchronous replicate system. For outback catastrophe recovery systems, this mirror may not be done by integrated systems but plainly by extra applications on primary and secondary machines .
extra benefits [edit ]
In addition to providing an extra replicate of the datum for the aim of redundancy in case of hardware failure, magnetic disk mirror can allow each phonograph record to be accessed individually for reading purposes. Under certain circumstances, this can significantly improve performance as the system can choose for each read which phonograph record can seek most cursorily to the required datum. This is specially significant where there are respective tasks competing for data on the same magnetic disk, and convulse ( where the switch between tasks takes up more clock than the job itself ) can be reduced. This is an important consideration in hardware configurations that frequently access the datum on the disk. In some implementations, the mirror disk can be split off and used for data backing, allowing the first gear phonograph record to remain active. however merging the two disks then may require a synchronization time period if any write I/O bodily process has occurred to the mirror magnetic disk.
early schemes [edit ]
Some mirror schemes employ three disks, with two of the disks for the redundancy mirror and the one-third to be split off for performing backups. In EMC terminology, these one-third disks are called business duration volumes ( BCVs ) .