File system dump/restore by node numbering
View Patent ↗A method and system for replicating a file system organized as a data structure from a source server to a target server. The replication process preserves the order of the inode table entries from the source file system to in the target file system at all stages of the replication process, as well as the internal consistency of the replica of the data structure on the target server. This replication eliminates use and employment of a map as the order of the inode index is preserved during the replication process.
1. A method for replicating a file system comprising:
creating an empty temporary inode table at a target file system;
scanning a source file system organized as a data structure in inode order, including for each object in a directory recording a pairlist with a second field associated with an object name and a first field associated with an object inode number in the source file system;
identifying an inode number of each directory in the source file system and entering each directory inode number in the temporary table;
creating a new entry order of directory entries in the temporary table, including sorting the temporary table entries according to the pairlist for each table entry, wherein a directory entry in the sorted temporary table precedes an entry for an object in said directory;
creating directories in the target system according to the new order, each directory created at a same inode number as in the source file system; and
copying all objects in a directory from the source file system to the target file system, wherein each object is copied at an inode number identified in the first field of a pairlist associated with said object, and wherein the directories are created before contents of the directories are copied.
2. The method of claim 1 , further comprising preserving inode number of each object from said source to said target file system at all stages of replication of directory indices.
3. The method of claim 1 , wherein remnants of said target system are salvageable.
4. A computer system comprising:
a processor coupled to memory, the processor operable to perform operations comprising:
creating an empty temporary inode table at a target file system;
scanning a source file system organized as a data structure in inode order, including for each object in a directory recording a pairlist with a second field associated with an object name and a first field associated with an object inode number in the source file system;
identifying an inode number of each directory in the source file system and entering each directory inode number in the temporary table;
creating a new entry order of directory entries in the temporary table, including sorting the temporary table entries according to the pairlist for each table entry, wherein a directory entry in the sorted temporary table precedes an entry for an object in said directory;
creating directories and files in the target system according to the new order, each directory created at a same inode number as in the source file system; and
copying all objects in a directory from the source file system to the target file system, wherein each object is copied at an inode number identified in the first field of a pairlist associated with said object, and wherein the directories are created before contents of the directories are copied.
5. The system of claim 4 , further comprising instructions to preserve inode number of each object from said source to said target file system at all stages of replication of directory indices.
6. The method of claim 4 , wherein remnants of said target system are salvageable.