Michael Renner wrote:
> Hi,
>
> this is a small update to the first paragraph of the WAL configuration
> chapter, going into more detail WRT redo vs. checkpoint records, since
> the underlying behavior is currently only deducible from the source. I'm
> not perfectly sure if I got everything right, so feel free to change as
> necessary.
>
> I think it'd be more appropriate to split the chapter and separate
> basics from implementation details and tuneables, but for time being
> this ought to suffice. Is somebody "in charge" of the documentation and
> overall structure or is it a community effort as everything else?
>
I read over you patch and I was afraid it was trying to put too much
information into a single paragraph, so I added a second paragraph that
just talks about checkpoint smoothing. I did not address the issue of
when the REDO WAL entry is written --- that is probably too much detail
for our documentation.
New patch attached, and applied.
---------------------------------------------------------------------------
>
> Best regards,
> Michael Renner
> diff --git a/doc/src/sgml/wal.sgml b/doc/src/sgml/wal.sgml
> index cff6fde..69b8b0a 100644
> --- a/doc/src/sgml/wal.sgml
> +++ b/doc/src/sgml/wal.sgml
> @@ -322,19 +322,24 @@
> </para>
>
> <para>
> - <firstterm>Checkpoints</firstterm><indexterm><primary>checkpoint</></>
> - are points in the sequence of transactions at which it is guaranteed
> - that the data files have been updated with all information written 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.
> - In the event of a crash, the crash recovery procedure looks at the latest
> - checkpoint record to determine the point in the log (known as the redo
> - record) from which it should start the REDO operation. Any changes made to
> - data files before that point are known to be already on disk. Hence, after
> - a checkpoint has been made, any log segments preceding the one containing
> - the redo record are no longer needed and can be recycled or removed. (When
> - <acronym>WAL</acronym> archiving is being done, the log segments must be
> - archived before being recycled or removed.)
> + <firstterm>Checkpoints</firstterm><indexterm><primary>checkpoint</></> are
> + points in the logical sequence of transactions at which it is guaranteed
> + that the data files have been updated with all information created before
> + the start of the checkpoint command. Since flushing all dirty data (meaning
> + "changed only in the WAL") to disk can take a while on databases with
> + write-heavy loads, checkpoints are not a single operation but rather a
> + series of events. When a checkpoint starts, a redo record is written to the
> + WAL and PostgreSQL starts writing out dirty data which has accumulated up to
> + the redo record. At checkpoint completion time, all changed files are
> + fsynced and a special checkpoint record is written to the log file. In the
> + event of a crash, the crash recovery procedure looks at the latest
> + checkpoint record to determine from which redo record it should start the
> + REDO operation. Any changes made to data files before that point are known
> + to be already on disk. Hence, after a checkpoint has been made, any log
> + segments preceding the one containing the redo record are no longer needed
> + and can be recycled or removed. (When <acronym>WAL</acronym> archiving is
> + being done, the log segments must be archived before being recycled or
> + removed.)
> </para>
>
> <para>
>
> --
> Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-hackers
--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ If your life is a hard drive, Christ can be your backup. +
Index: doc/src/sgml/wal.sgml
===================================================================
RCS file: /cvsroot/pgsql/doc/src/sgml/wal.sgml,v
retrieving revision 1.58
diff -c -c -r1.58 wal.sgml
*** doc/src/sgml/wal.sgml 15 Jan 2009 00:34:25 -0000 1.58
--- doc/src/sgml/wal.sgml 9 Apr 2009 16:19:18 -0000
***************
*** 326,343 ****
are points in the sequence of transactions at which it is guaranteed
that the data files have been updated with all information written 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.
In the event of a crash, the crash recovery procedure looks at the latest
checkpoint record to determine the point in the log (known as the redo
record) from which it should start the REDO operation. Any changes made to
! data files before that point are known to be already on disk. Hence, after
! a checkpoint has been made, any log segments preceding the one containing
the redo record are no longer needed and can be recycled or removed. (When
<acronym>WAL</acronym> archiving is being done, the log segments must be
archived before being recycled or removed.)
</para>
<para>
The server's background writer process will automatically perform
a checkpoint every so often. A checkpoint is created every <xref
linkend="guc-checkpoint-segments"> log segments, or every <xref
--- 326,352 ----
are points in the sequence of transactions at which it is guaranteed
that the data files have been updated with all information written 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.
! (The changes were previously flushed to the <acronym>WAL</acronym> files.)
In the event of a crash, the crash recovery procedure looks at the latest
checkpoint record to determine the point in the log (known as the redo
record) from which it should start the REDO operation. Any changes made to
! data files before that point are guaranteed to be already on disk. Hence, after
! a checkpoint, log segments preceding the one containing
the redo record are no longer needed and can be recycled or removed. (When
<acronym>WAL</acronym> archiving is being done, the log segments must be
archived before being recycled or removed.)
</para>
<para>
+ The checkpoint requirement of flushing all dirty data pages to disk
+ can cause a significant I/O load. For this reason, checkpoint
+ activity is throttled so I/O begins at checkpoint start and completes
+ before the next checkpoint starts; this minimizes performance
+ degradation during checkpoints.
+ </para>
+
+ <para>
The server's background writer process will automatically perform
a checkpoint every so often. A checkpoint is created every <xref
linkend="guc-checkpoint-segments"> log segments, or every <xref
