On Wed, Sep 29, 2021 at 4:07 PM Thomas Munro <thomas.mu...@gmail.com> wrote:
> Hmm, on closer inspection, isn't the lack of real interlocking with
> checkpoints a bit suspect already? What stops bgwriter from writing
> to the previous relfilenode generation's fd, if a relfilenode is
> recycled while BgBufferSync() is running? Not sinval, and not the
> above code that only runs between BgBufferSync() invocations.
I managed to produce a case where live data is written to an unlinked
file and lost, with a couple of tweaks to get the right timing and
simulate OID wraparound. See attached. If you run the following
commands repeatedly with shared_buffers=256kB and
bgwriter_lru_multiplier=10, you should see a number lower than 10,000
from the last query in some runs, depending on timing.
create extension if not exists chaos;
create extension if not exists pg_prewarm;
drop table if exists t1, t2;
checkpoint;
vacuum pg_class;
select clobber_next_oid(200000);
create table t1 as select 42 i from generate_series(1, 10000);
select pg_prewarm('t1'); -- fill buffer pool with t1
update t1 set i = i; -- dirty t1 buffers so bgwriter writes some
select pg_sleep(2); -- give bgwriter some time
drop table t1;
checkpoint;
vacuum pg_class;
select clobber_next_oid(200000);
create table t2 as select 0 i from generate_series(1, 10000);
select pg_prewarm('t2'); -- fill buffer pool with t2
update t2 set i = 1 where i = 0; -- dirty t2 buffers so bgwriter writes some
select pg_sleep(2); -- give bgwriter some time
select pg_prewarm('pg_attribute'); -- evict all clean t2 buffers
select sum(i) as t2_sum_should_be_10000 from t2; -- have any updates been lost?
From b116b80b2775b004e35a9e7be0a057ee2724041b Mon Sep 17 00:00:00 2001
From: Thomas Munro <thomas.mu...@gmail.com>
Date: Thu, 30 Sep 2021 15:47:23 +1300
Subject: [PATCH 1/2] HACK: A function to control the OID allocator.
---
src/test/modules/chaos/Makefile | 21 +++++++++++++++++++++
src/test/modules/chaos/chaos--1.0.sql | 8 ++++++++
src/test/modules/chaos/chaos.c | 26 ++++++++++++++++++++++++++
src/test/modules/chaos/chaos.control | 4 ++++
4 files changed, 59 insertions(+)
create mode 100644 src/test/modules/chaos/Makefile
create mode 100644 src/test/modules/chaos/chaos--1.0.sql
create mode 100644 src/test/modules/chaos/chaos.c
create mode 100644 src/test/modules/chaos/chaos.control
diff --git a/src/test/modules/chaos/Makefile b/src/test/modules/chaos/Makefile
new file mode 100644
index 0000000000..ac69721af6
--- /dev/null
+++ b/src/test/modules/chaos/Makefile
@@ -0,0 +1,21 @@
+# src/test/modules/chaos/Makefile
+
+MODULE_big = chaos
+OBJS = \
+ $(WIN32RES) \
+ chaos.o
+PGFILEDESC = "chaos - module in which to write throwaway fault-injection hacks"
+
+EXTENSION = chaos
+DATA = chaos--1.0.sql
+
+ifdef USE_PGXS
+PG_CONFIG = pg_config
+PGXS := $(shell $(PG_CONFIG) --pgxs)
+include $(PGXS)
+else
+subdir = src/test/modules/chaos
+top_builddir = ../../../..
+include $(top_builddir)/src/Makefile.global
+include $(top_srcdir)/contrib/contrib-global.mk
+endif
diff --git a/src/test/modules/chaos/chaos--1.0.sql b/src/test/modules/chaos/chaos--1.0.sql
new file mode 100644
index 0000000000..5016f7e586
--- /dev/null
+++ b/src/test/modules/chaos/chaos--1.0.sql
@@ -0,0 +1,8 @@
+/* src/test/modules/chaos/chaos--1.0.sql */
+
+-- complain if script is sourced in psql, rather than via CREATE EXTENSION
+\echo Use "CREATE EXTENSION chaos" to load this file. \quit
+
+CREATE FUNCTION clobber_next_oid(size BIGINT)
+ RETURNS pg_catalog.void STRICT
+ AS 'MODULE_PATHNAME' LANGUAGE C;
diff --git a/src/test/modules/chaos/chaos.c b/src/test/modules/chaos/chaos.c
new file mode 100644
index 0000000000..f1052f865e
--- /dev/null
+++ b/src/test/modules/chaos/chaos.c
@@ -0,0 +1,26 @@
+#include "postgres.h"
+
+#include "access/transam.h"
+#include "fmgr.h"
+#include "storage/lwlock.h"
+
+#include <limits.h>
+
+PG_MODULE_MAGIC;
+
+PG_FUNCTION_INFO_V1(clobber_next_oid);
+
+Datum
+clobber_next_oid(PG_FUNCTION_ARGS)
+{
+ int64 oid = PG_GETARG_INT64(0);
+
+ if (oid < FirstNormalObjectId || oid > UINT_MAX)
+ elog(ERROR, "invalid oid");
+
+ LWLockAcquire(OidGenLock, LW_EXCLUSIVE);
+ ShmemVariableCache->nextOid = oid;
+ LWLockRelease(OidGenLock);
+
+ PG_RETURN_VOID();
+}
diff --git a/src/test/modules/chaos/chaos.control b/src/test/modules/chaos/chaos.control
new file mode 100644
index 0000000000..137ab8a58d
--- /dev/null
+++ b/src/test/modules/chaos/chaos.control
@@ -0,0 +1,4 @@
+comment = 'Test module for throwaway fault-injection hacks...'
+default_version = '1.0'
+module_pathname = '$libdir/chaos'
+relocatable = true
--
2.30.2
From 2acc2ad31c1db268de0e8927d5c10ba2bb06e33c Mon Sep 17 00:00:00 2001
From: Thomas Munro <thomas.mu...@gmail.com>
Date: Thu, 30 Sep 2021 17:16:01 +1300
Subject: [PATCH 2/2] HACK: Slow the bgwriter down a bit.
---
src/backend/postmaster/bgwriter.c | 2 ++
src/backend/storage/buffer/bufmgr.c | 2 ++
2 files changed, 4 insertions(+)
diff --git a/src/backend/postmaster/bgwriter.c b/src/backend/postmaster/bgwriter.c
index 5584f4bc24..b65284b1f6 100644
--- a/src/backend/postmaster/bgwriter.c
+++ b/src/backend/postmaster/bgwriter.c
@@ -238,7 +238,9 @@ BackgroundWriterMain(void)
/*
* Do one cycle of dirty-buffer writing.
*/
+ elog(LOG, "=== begin BgBufferSync ===");
can_hibernate = BgBufferSync(&wb_context);
+ elog(LOG, "=== end BgBufferSync ===");
/*
* Send off activity statistics to the stats collector
diff --git a/src/backend/storage/buffer/bufmgr.c b/src/backend/storage/buffer/bufmgr.c
index e88e4e918b..989125e37f 100644
--- a/src/backend/storage/buffer/bufmgr.c
+++ b/src/backend/storage/buffer/bufmgr.c
@@ -2452,6 +2452,8 @@ BgBufferSync(WritebackContext *wb_context)
}
else if (sync_state & BUF_REUSABLE)
reusable_buffers++;
+
+ pg_usleep(1000000);
}
PendingBgWriterStats.m_buf_written_clean += num_written;
--
2.30.2
