Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commitcaa0783

Browse files
committed
Improve handling and coverage of --no-ensure-shutdown in pg_rewind
This includes a couple of changes around the new behavior of pg_rewindwhich enforces recovery to happen once on a cluster not shut downcleanly:- Some comments and documentation improvements.- Shutdown the cluster to rewind with immediate mode in all the tests,this allows to check after the forced recovery behavior which is wantedas new default.- Use -F for the forced recovery step, so as postgres does not usefsync. This was useless as a final sync is done once the tool is done.Author: Michael PaquierReviewed-by: Alexey KondratovDiscussion:https://postgr.es/m/20191004083721.GA1829@paquier.xyz
1 parent732457b commitcaa0783

File tree

4 files changed

+34
-19
lines changed

4 files changed

+34
-19
lines changed

‎doc/src/sgml/ref/pg_rewind.sgml

Lines changed: 6 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -169,12 +169,14 @@ PostgreSQL documentation
169169
<term><option>--no-ensure-shutdown</option></term>
170170
<listitem>
171171
<para>
172-
<application>pg_rewind</application> verifies that the target server
173-
is cleanly shutdown before rewinding; by default, if it isn't, it
174-
starts the server in single-user mode to complete crash recovery.
172+
<application>pg_rewind</application> requires that the target server
173+
is cleanly shut down before rewinding. By default, if the target server
174+
is not shut down cleanly, <application>pg_rewind</application> starts
175+
the target server in single-user mode to complete crash recovery first,
176+
and stops it.
175177
By passing this option, <application>pg_rewind</application> skips
176178
this and errors out immediately if the server is not cleanly shut
177-
down.Users are expected to handle the situation themselves in that
179+
down. Users are expected to handle the situation themselves in that
178180
case.
179181
</para>
180182
</listitem>

‎src/bin/pg_rewind/pg_rewind.c

Lines changed: 12 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -270,11 +270,12 @@ main(int argc, char **argv)
270270
pg_free(buffer);
271271

272272
/*
273-
* If the target instance was not cleanly shut down, run a single-user
274-
* postgres session really quickly and reload the control file to get the
275-
* new state. Note if no_ensure_shutdown is specified, pg_rewind won't do
276-
* that automatically. That means users need to do themselves in advance,
277-
* else pg_rewind will soon quit, see sanityChecks().
273+
* If the target instance was not cleanly shut down, start and stop the
274+
* target cluster once in single-user mode to enforce recovery to finish,
275+
* ensuring that the cluster can be used by pg_rewind. Note that if
276+
* no_ensure_shutdown is specified, pg_rewind ignores this step, and users
277+
* need to make sure by themselves that the target cluster is in a clean
278+
* state.
278279
*/
279280
if (!no_ensure_shutdown&&
280281
ControlFile_target.state!=DB_SHUTDOWNED&&
@@ -847,8 +848,12 @@ ensureCleanShutdown(const char *argv0)
847848
if (dry_run)
848849
return;
849850

850-
/* finally run postgres in single-user mode */
851-
snprintf(cmd,MAXCMDLEN,"\"%s\" --single -D \"%s\" template1 < \"%s\"",
851+
/*
852+
* Finally run postgres in single-user mode. There is no need to use
853+
* fsync here. This makes the recovery faster, and the target data folder
854+
* is synced at the end anyway.
855+
*/
856+
snprintf(cmd,MAXCMDLEN,"\"%s\" --single -F -D \"%s\" template1 < \"%s\"",
852857
exec_path,datadir_target,DEVNULL);
853858

854859
if (system(cmd)!=0)

‎src/bin/pg_rewind/t/005_same_timeline.pl

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,7 @@
1+
#
2+
# Test that running pg_rewind with the source and target clusters
3+
# on the same timeline runs successfully.
4+
#
15
use strict;
26
use warnings;
37
use TestLib;

‎src/bin/pg_rewind/t/RewindTest.pm

Lines changed: 12 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -227,8 +227,10 @@ sub run_pg_rewind
227227
# Append the rewind-specific role to the connection string.
228228
$standby_connstr ="$standby_connstr user=rewind_user";
229229

230-
# Stop the master and be ready to perform the rewind
231-
$node_master->stop;
230+
# Stop the master and be ready to perform the rewind. The cluster
231+
# needs recovery to finish once, and pg_rewind makes sure that it
232+
# happens automatically.
233+
$node_master->stop('immediate');
232234

233235
# At this point, the rewind processing is ready to run.
234236
# We now have a very simple scenario with a few diverged WAL record.
@@ -260,19 +262,21 @@ sub run_pg_rewind
260262
}
261263
elsif ($test_modeeq"remote")
262264
{
263-
264-
#Do rewind using a remote connection as source
265+
# Do rewind using a remote connection as source, generating
266+
#recovery configuration automatically.
265267
command_ok(
266268
[
267269
'pg_rewind',"--debug",
268270
"--source-server",$standby_connstr,
269-
"--target-pgdata=$master_pgdata","-R",
270-
"--no-sync"
271+
"--target-pgdata=$master_pgdata","--no-sync",
272+
"--write-recovery-conf"
271273
],
272274
'pg_rewind remote');
273275

274-
# Check that standby.signal has been created.
275-
ok(-e"$master_pgdata/standby.signal");
276+
# Check that standby.signal is here as recovery configuration
277+
# was requested.
278+
ok(-e"$master_pgdata/standby.signal",
279+
'standby.signal created after pg_rewind');
276280

277281
# Now, when pg_rewind apparently succeeded with minimal permissions,
278282
# add REPLICATION privilege. So we could test that new standby

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp