How to perform recovery using
nonquiecent checkpointing using UNDO logging
After a
system failure,
the UNDO log file can be
in one of
2 possible states:
When scanning the
UNDO log file
backwards,
we find a
<END CKPT> record
first
How to perform recovery using
nonquiecent checkpointing using UNDO logging
After a
system failure,
the UNDO log file can be
in one of
2 possible states:
When scanning the
UNDO log file
backwards,
we find a
<START CKPT> record
first
UNDO logging recovery procedure for
case 1 (completed checkpoint)
Case 1 scenario:
I.e.:
the last log checkpoint operation
has
completed successfull
UNDO logging recovery procedure for
case 1 (completed checkpoint)
Case 1 scenario:
Because the portion of the
UNDO log
before the
<START CKPT> record contain
UNDO actions of
T1, T2, ...,
Tk
that we know have
complete successfull
-
they are no longer needed !
UNDO logging recovery procedure for
case 1 (completed checkpoint)
Case 1 scenario:
Because
some of
these transaction may
be incomplete
-
we must clean up
partial updates made by these
incomplete transactions
UNDO logging recovery procedure for
case 1 (completed checkpoint)
Recovery procedure for
case 1 scenario:
- Step 1:
This location in the
log file is the
last location that we need to
search for
UNDO records
UNDO logging recovery procedure for
case 1 (completed checkpoint)
Recovery procedure for
case 1 scenario:
- Step 2:
UNDO logging recovery procedure for
case 1 (completed checkpoint)
Recovery procedure for
case 1 scenario:
- Step 3:
UNDO logging recovery procedure for
case 1
- Example recovery
Example recovery for
case 1 scenario:
UNDO logging recovery procedure for
case 1
- Example recovery
Example recovery for
case 1 scenario:
UNDO logging recovery procedure for
case 1
- Example recovery
Example recovery for
case 1 scenario:
Click on
figure to
pull out...
UNDO logging recovery procedure for
case 1
- Example recovery
Example recovery for
case 1 scenario:
UNDO logging recovery procedure for
case 2 (crashed during checkpointing)
Case 2 scenario:
I.e.:
the system
crashed
during
a
log checkpoint operation
UNDO logging recovery procedure for
case 2 (crashed during checkpointing)
Case 2 scenario:
I.e.:
scan further back to
find the <START T1> record
--- this is the
earliest record that we need to use.
UNDO logging recovery procedure for
case 2 (crashed during checkpointing)
Case 2 scenario:
Again:
we must undo the
changes made by
the incomplete transactions
(in the given UNDO log portion)
UNDO logging recovery procedure for
case 2 (crashed during checkpointing)
Recovery procedure for
case 2 scenario:
- Step 1:
This
location
in the
UNDO log file
log file is the last location
that we need to
search for
UNDO records
UNDO logging recovery procedure for
case 2 (crashed during checkpointing)
Recovery procedure for
case 2 scenario:
- Step 2:
UNDO logging recovery procedure for
case 2 (crashed during checkpointing)
Recovery procedure for
case 2 scenario:
- Step 2:
UNDO logging recovery procedure for
case 2
- Example recovery
Example recovery for
case 2 scenario:
UNDO logging recovery procedure for
case 2
- Example recovery
Example recovery for
case 2 scenario:
UNDO logging recovery procedure for
case 2
- Example recovery
Example recovery for
case 2 scenario:
Click on
figure
to pull out...
UNDO logging recovery procedure for
case 2
- Example recovery
Example recovery for
case 2 scenario:
❮
❯