Re: [PATCH]make pg_rewind to not copy useless WAL files

On 2017-09-30 02:17:54，"Alexander Korotkov" <a(dot)korotkov(at)postgrespro(dot)ru> wrote:

Great. Now code of this patch looks good for me.
However, we forgot about documentation.

<para>
The result is equivalent to replacing the target data directory with the
source one. Only changed blocks from relation files are copied;
all other files are copied in full, including configuration files. The
advantage of <application>pg_rewind</> over taking a new base backup, or
tools like <application>rsync</>, is that <application>pg_rewind</> does
not require reading through unchanged blocks in the cluster. This makes
it a lot faster when the database is large and only a small
fraction of blocks differ between the clusters.
</para>

At least, this paragraph need to be adjusted, because it states whose files are copied. And probably latter paragraphs whose state about WAL files.

Your are rigth.
I wrote a draft as following, but i'm afraid whether the english statement is accurate.

----------------------
diff --git a/doc/src/sgml/ref/pg_rewind.sgml b/doc/src/sgml/ref/pg_rewind.sgml
index d5430d4..bcd094b 100644
--- a/doc/src/sgml/ref/pg_rewind.sgml
+++ b/doc/src/sgml/ref/pg_rewind.sgml
@@ -50,9 +50,12 @@ PostgreSQL documentation
<para>
The result is equivalent to replacing the target data directory with the
source one. Only changed blocks from relation files are copied;
- all other files are copied in full, including configuration files. The
- advantage of <application>pg_rewind</> over taking a new base backup, or
- tools like <application>rsync</>, is that <application>pg_rewind</> does
+ all other files except WAL are copied in full, including configuration
+ files. Only the WAL files between the point of divergence and the current
+ WAL insert location of the source server are copied, for other WAL files
+ are useless for the target server. The advantage of
+ <application>pg_rewind</> over taking a new base backup, or tools
+ like <application>rsync</>, is that <application>pg_rewind</> does
not require reading through unchanged blocks in the cluster. This makes
it a lot faster when the database is large and only a small
fraction of blocks differ between the clusters.
@@ -231,7 +234,7 @@ PostgreSQL documentation
<para>
Copy all other files such as <filename>pg_xact</filename> and
configuration files from the source cluster to the target cluster
- (everything except the relation files).
+ (everything except the relation files and some WAL files).
</para>
</step>
<step>

------
Best Regars,
Chen Huajun