Write-Ahead Logging (WAL)
Write-Ahead Logging (WAL)
is a standard approach to transaction logging. Its detailed
description may be found in most (if not all) books about
transaction processing. Briefly, WAL's central
concept is that changes to data files (where tables and indexes
reside) must be written only after those changes have been logged,
that is, when log records have been flushed to permanent
storage. If we follow this procedure, we do not need to flush
data pages to disk on every transaction commit, because we know
that in the event of a crash we will be able to recover the
database using the log: any changes that have not been applied to
the data pages will first be redone from the log records (this is
roll-forward recovery, also known as REDO) and then changes made by
uncommitted transactions will be removed from the data pages
(roll-backward recovery, UNDO).
Benefits of WAL
The first obvious benefit of using WAL is a
significantly reduced number of disk writes, since only the log
file needs to be flushed to disk at the time of transaction
commit; in multiuser environments, commits of many transactions
may be accomplished with a single fsync() of
the log file. Furthermore, the log file is written sequentially,
and so the cost of syncing the log is much less than the cost of
flushing the data pages.
The next benefit is consistency of the data pages. The truth is
that, before WAL,
PostgreSQL was never able to guarantee
consistency in the case of a crash. Before
WAL, any crash during writing could result in:
index rows pointing to nonexistent table rows
index rows lost in split operations
totally corrupted table or index page content, because
of partially written data pages
Problems with indexes (problems 1 and 2) could possibly have been
fixed by additional fsync() calls, but it is
not obvious how to handle the last case without
WAL; WAL saves the entire data
page content in the log if that is required to ensure page
consistency for after-crash recovery.
Future Benefits
The UNDO operation is not implemented. This means that changes
made by aborted transactions will still occupy disk space and that
a permanent pg_clog file to hold
the status of transactions is still needed, since
transaction identifiers cannot be reused. Once UNDO is implemented,
pg_clog will no longer be required to be
permanent; it will be possible to remove
pg_clog at shutdown. (However, the urgency of
this concern has decreased greatly with the adoption of a segmented
storage method for pg_clog: it is no longer
necessary to keep old pg_clog entries around
forever.)
With UNDO, it will also be possible to implement
savepoints to allow partial rollback of
invalid transaction operations (parser errors caused by mistyping
commands, insertion of duplicate primary/unique keys and so on)
with the ability to continue or commit valid operations made by
the transaction before the error. At present, any error will
invalidate the whole transaction and require a transaction abort.
WAL offers the opportunity for a new method for
database on-line backup and restore (BAR). To
use this method, one would have to make periodic saves of data
files to another disk, a tape or another host and also archive the
WAL log files. The database file copy and the
archived log files could be used to restore just as if one were
restoring after a crash. Each time a new database file copy was
made the old log files could be removed. Implementing this
facility will require the logging of data file and index creation
and deletion; it will also require development of a method for
copying the data files (operating system copy commands are not
suitable).
A difficulty standing in the way of realizing these benefits is that
they require saving WAL entries for considerable
periods of time (e.g., as long as the longest possible transaction if
transaction UNDO is wanted). The present WAL
format is extremely bulky since it includes many disk page
snapshots. This is not a serious concern at present, since the
entries only need to be kept for one or two checkpoint intervals;
but to achieve these future benefits some sort of compressed
WAL format will be needed.
WAL Configuration
There are several WAL-related configuration parameters that
affect database performance. This section explains their use.
Consult for details about setting
configuration parameters.
Checkpoints are points in the sequence of
transactions at which it is guaranteed that the data files have
been updated with all information logged before the checkpoint. At
checkpoint time, all dirty data pages are flushed to disk and a
special checkpoint record is written to the log file. As result, in
the event of a crash, the recoverer knows from what record in the
log (known as the redo record) it should start the REDO operation,
since any changes made to data files before that record are already
on disk. After a checkpoint has been made, any log segments written
before the redo records are no longer needed and can be recycled or
removed. (When WAL-based BAR is
implemented, the log segments would be archived before being recycled
or removed.)
The server spawns a special process every so often
to create the next checkpoint. A checkpoint is created every
checkpoint_segments log segments, or every
checkpoint_timeout seconds, whichever comes first.
The default settings are 3 segments and 300 seconds respectively.
It is also possible to force a checkpoint by using the SQL command
CHECKPOINT.
Reducing checkpoint_segments and/or
checkpoint_timeout causes checkpoints to be done
more often. This allows faster after-crash recovery (since less work
will need to be redone). However, one must balance this against the
increased cost of flushing dirty data pages more often. In addition,
to ensure data page consistency, the first modification of a data
page after each checkpoint results in logging the entire page
content. Thus a smaller checkpoint interval increases the volume of
output to the log, partially negating the goal of using a smaller
interval, and in any case causing more disk I/O.
There will be at least one 16 MB segment file, and will normally
not be more than 2 * checkpoint_segments + 1
files. You can use this to estimate space requirements for WAL.
Ordinarily, when old log segment files are no longer needed, they
are recycled (renamed to become the next segments in the numbered
sequence). If, due to a short-term peak of log output rate, there
are more than 2 * checkpoint_segments + 1
segment files, the unneeded segment files will be deleted instead
of recycled until the system gets back under this limit.
There are two commonly used WAL functions:
LogInsert and LogFlush.
LogInsert is used to place a new record into
the WAL buffers in shared memory. If there is no
space for the new record, LogInsert will have
to write (move to kernel cache) a few filled WAL
buffers. This is undesirable because LogInsert
is used on every database low level modification (for example,
row insertion) at a time when an exclusive lock is held on
affected data pages, so the operation needs to be as fast as
possible. What is worse, writing WAL buffers may
also force the creation of a new log segment, which takes even more
time. Normally, WAL buffers should be written
and flushed by a LogFlush request, which is
made, for the most part, at transaction commit time to ensure that
transaction records are flushed to permanent storage. On systems
with high log output, LogFlush requests may
not occur often enough to prevent WAL buffers
being written by LogInsert. On such systems
one should increase the number of WAL buffers by
modifying the configuration parameter wal_buffers.
The default number of
WAL buffers is 8. Increasing this value will
correspondingly increase shared memory usage.
Checkpoints are fairly expensive because they force all dirty kernel
buffers to disk using the operating system sync()> call.
Busy servers may fill checkpoint segment files too quickly,
causing excessive checkpointing. If such forced checkpoints happen
more frequently than checkpoint_warning seconds,
a message, will be output to the server logs recommending increasing
checkpoint_segments.
The commit_delay parameter defines for how many
microseconds the server process will sleep after writing a commit
record to the log with LogInsert but before
performing a LogFlush. This delay allows other
server processes to add their commit records to the log so as to have all
of them flushed with a single log sync. No sleep will occur if fsync
is not enabled or if fewer than commit_siblings
other sessons are currently in active transactions; this avoids
sleeping when it's unlikely that any other session will commit soon.
Note that on most platforms, the resolution of a sleep request is
ten milliseconds, so that any nonzero commit_delay
setting between 1 and 10000 microseconds would have the same effect.
Good values for these parameters are not yet clear; experimentation
is encouraged.
The wal_sync_method parameter determines how
PostgreSQL will ask the kernel to force
WAL updates out to disk.
All the options should be the same as far as reliability goes,
but it's quite platform-specific which one will be the fastest.
Note that this parameter is irrelevant if fsync
has been turned off.
Setting the wal_debug parameter to any nonzero
value will result in each LogInsert and
LogFlush WAL call being
logged to the server log. At present, it makes no difference what
the nonzero value is. This option may be replaced by a more
general mechanism in the future.
Internals
WAL is automatically enabled; no action is
required from the administrator except ensuring that the additional
disk-space requirements of the WAL logs are met,
and that any necessary tuning is done (see ).
WAL logs are stored in the directory
pg_xlog under the data directory, as a set of
segment files, each 16 MB in size. Each segment is divided into 8
kB pages. The log record headers are described in
access/xlog.h; the record content is dependent
on the type of event that is being logged. Segment files are given
ever-increasing numbers as names, starting at
0000000000000000. The numbers do not wrap, at
present, but it should take a very long time to exhaust the
available stock of numbers.
The WAL buffers and control structure are in
shared memory and are handled by the server child processes; they
are protected by lightweight locks. The demand on shared memory is
dependent on the number of buffers. The default size of the
WAL buffers is 8 buffers of 8 kB each, or 64 kB
total.
It is of advantage if the log is located on another disk than the
main database files. This may be achieved by moving the directory
pg_xlog to another location (while the server
is shut down, of course) and creating a symbolic link from the
original location in the main data directory to the new location.
The aim of WAL, to ensure that the log is
written before database records are altered, may be subverted by
disk drives that falsely report a successful write to the kernel,
when, in fact, they have only cached the data and not yet stored it
on the disk. A power failure in such a situation may still lead to
irrecoverable data corruption. Administrators should try to ensure
that disks holding PostgreSQL's
WAL log files do not make such false reports.
After a checkpoint has been made and the log flushed, the
checkpoint's position is saved in the file
pg_control. Therefore, when recovery is to be
done, the server first reads pg_control and
then the checkpoint record; then it performs the REDO operation by
scanning forward from the log position indicated in the checkpoint
record. Because the entire content of data pages is saved in the
log on the first page modification after a checkpoint, all pages
changed since the checkpoint will be restored to a consistent
state.
Using pg_control to get the checkpoint
position speeds up the recovery process, but to handle possible
corruption of pg_control, we should actually
implement the reading of existing log segments in reverse order --
newest to oldest -- in order to find the last checkpoint. This has
not been implemented, yet.