Method for faster recovery of a potentially corrupted temporary file system.
Original Publication Date: 2004-Mar-12
Included in the Prior Art Database: 2004-Mar-12
Method for faster recovery of a potentially corrupted temporary file system by rebuilding the file system in the event that the consistency of a file system is detected.
Method for faster recovery of a potentially corrupted temporary file system .
Disclosed is a method for decreasing the time necessary for recovery of potentially corrupted or inconsistent temporary file systems. Today if a journaled file system is dirty the file system log is replayed on mount. If the replay of the log is unsuccessful or this is not a journaled file system the file system meta data must be made consistent again with a utility such as fsck. A multiple terabyte file system can take a long time to fix with fsck. Some applications can generate several terabytes of information that is only valid during the current invocation of the application so fixing a potentially inconsistent file system is an expensive operation for these applications.
This disclosure proposes enhancements to the file system layer of an OS such that a file system can be marked as a volatile file system such that if the replaying of the log fails the mount helper for the file system will automatically recreate the file system. This disclose proposes this action because the time to create a multi-terabtye extent based file system can be significantly less than the time it takes to run fsck. The file system could support a number of modes for a volatile file system. In one mode the file system would be recreated by the file system mount helper in the event the file system is marked as mounted, in another the file system would be recreated if the mount helper detected that the log an...