| 
|
Topic: log-structured rollback-recovery
topics >
computer science >
backtracking
Topic:
database consistency and reliability
Topic:
database implementation
Topic:
database transactions
Topic:
error safe systems
Topic:
file system reliability
Topic:
log-structured file system
Topic:
logging data and events
Topic:
reliability of distributed systems
Topic:
resourceful, redundant systems for reliability
Topic:
updating information in a distributed system
Summary
Logs allow rollback and recovery of failed transactions and failed nodes. All updates must be logged. Accesses need logging if they can effect updates. Periodic checkpoints may simplify recovery. (cbb 8/06)
Subtopic: update log, differential file
Quote: log and checkpoint memory-resident data using a destructive checkpoint; recover data from the checkpoint file Quote: log-based rollback-recovery must log all nondeterministic events with enough information to replay the event; useful for applications that frequently interact with the outside world [»elnoEN9_2002] Quote: signals and interrupts do not work with many log-based, rollback-recovery protocols; they require piecewise determinism between messages [»slyeJH10_1998] Quote: need other methods for deterministic replay of system calls and external inputs [»slyeJH10_1998] Quote: use software counters to allow deterministic replay of interrupts for log-based recovery of shared memory access on uniprocessor; 6-8% slowdown [»slyeJH10_1998] Quote: count backward branches to replay asynchronous interrupts [»slyeJH10_1998] Quote: for reliability, log all updates to an object; since append-only, non-volatility is cheap [»lampBW10_1983] Quote: causal logging is as fast as optimistic logging while allowing each process to commit output independently; roll-back to most recent checkpoint; more complex [»elnoEN9_2002]
Quote: a differential file stores all modifications to a database with periodic merging [»maryF_1982] Quote: for an online database, recorded update transactions on a differential file and updated at night; better performance, reliability, backup and recovery [»gremLL9_1982] Quote: used differential file for updating a large administrative database; 30,000 to 60,000 transactions per day [»gremLL9_1982, OK] Quote: very large databases cannot use dump tapes and update logs because reading a 1-terabyte dump will take days [»silbA10_1991] Quote: use logical logging for one-phase commit; reexecute log on failures [»abdaM12_1998]
Topic: backtracking (30 items)
Topic: database consistency and reliability (15 items)
Topic: database implementation (18 items)
Topic: database transactions (27 items)
Topic: error safe systems (76 items)
Topic: file system reliability (26 items)
Topic: log-structured file system (11 items)
Topic: logging data and events (17 items)
Topic: reliability of distributed systems (35 items)
Topic: resourceful, redundant systems for reliability (38 items)
Topic: updating information in a distributed system
(50 items)