AMASS Overview

File Recovery

Because write operations go to the cache first (including sync,

 

fsync, and synchronous IO), in the event of a system crash,

 

AMASS recovers all write operations that were in the cache and

 

processes the writes to media after the system is rebooted. This

 

paradigm provides a level of data integrity consistent with

 

standard UNIX file systems.

Database Recovery

Most UNIX file systems require fsck, a file system integrity

 

check, of all file systems mounted at the time of a system crash.

 

This checking can be very time-consuming. However, the

 

online File System Database, eliminates the need for this

 

checking. After you reboot the system, AMASS corrects its

 

Database based upon the /filesysdb/journal file (a

 

transaction log) and starts AMASS.

Library Recovery

Media can be left in a drive if a system crash occurs. AMASS

 

uses either external bar codes (network-attached libraries) or

 

internal headers (SCSI-attached libraries) to identify a volume

 

and automatically returns the “stranded” volume to its home

 

position.

 

Because these recovery functions are automated and can be run

 

from the startup script after a system reboot, the AMASS

 

recovery, startup, and file system mount can all be performed as

 

part of other UNIX file system operations.

Volume Verification

Volume verification is extremely important. Storage devices are

 

subject to operator error and automated systems can suffer from

 

hardware malfunctions leading to incorrect volumes being

 

loaded into the drives.

1-14

Online Archiving with AMASS

6-00026-01 Rev A