Restoring method and computer system

摘要

A NAS device providing a snapshot function for executing CoW processing in multiple successive block units is provided, wherein upon restoring a file saved in a file system sharing from a snapshot, the speed for restoring files is enhanced by restoring only data blocks that have been changed within the data blocks constituting a file, to thereby reduce the amount of data being restored and cut down CoW processing. Upon restoring a file stored in file system sharing from a snapshot file, data blocks storing different data between the restore source file and the restore destination file are specified based on a differential map table and a differential block table, and only the specified data blocks are restored. Further, upon restoring data from a file other than the snapshot, the data blocks constituting the restore source file and the restore destination file are compared, and only the data blocks storing different data are restored.

主权项

1. A restore method executed in a computer system having a file server and a storage system, wherein the storage system provides a first logical volume and a second logical volume to the file server; and
the method performed by the file server comprises:
a first step for saving data stored in a chunk composed of multiple successive blocks including a write destination block to the second logical volume upon writing data to the first logical volume, thereby writing data to the first logical volume constituting a file system sharing;a second step for managing blocks in which the first logical volume stores write data for each chunk;a third step for specifying the block having data written to the first logical volume out of the chunks being a restore target, wherein when restoring a file stored in the file system sharing so as to restore the first logical volume, the third step further includes specifying a block that differs from the block constituting a restore destination file out of blocks constituting a restore source file from a chunk saved in the second logical volume; anda fourth step for copying the saved data stored in the specified block to the first logical volume, wherein if the restore destination file does not have a fixed size, the third and fourth steps:
specify a block that differs from a block constituting the restore source file out of the blocks constituting the restore destination file;cancel the allocation of the specified block from the restore destination file;newly allocate to the restore destination file a number of blocks of the first logical volume corresponding to the number of blocks constituting the restore source file and that differs from the blocks constituting the restore destination file; andrestore the blocks constituting the restore source file to the newly allocated data blocks.