SFS Restart Recovery
Not all sudden failures of your computer system cause a data loss. When an SFS server is started, it determines whether anyone was making changes when it last ended. That is, SFS tries to find unfinished work. If it does, it automatically rolls back any uncommitted changes. By doing this, SFS automatically recovers from almost all system failures.
If your application also accesses other protected resources, this same protection is extended to those other resources through Coordinated Resource Recovery (CRR). For details on CRR, see Your Applications and Data Integrity.