pg /c database 通过。似乎是pg使用 open/write 写入了某个文件,通过 fopen/fscanf 绕过了hook路径导致的。出现了新的段错误。

This commit is contained in:
2026-03-13 10:49:31 +00:00
parent 544f532bf5
commit 4d350d5aea
28 changed files with 1686 additions and 128 deletions

View File

@@ -15,10 +15,14 @@ dynamic_shared_memory_type = sysv
# 4. 重启 PostgreSQL
## 清理旧空间
sudo systemctl stop postgresql
pkill -9 -u postgres postgres || true
rm -rf /home/lian/pg/pgdata
rm -rf /zvfs/pg_ts_bench
rm -rf /tmp/pg.log
## 初始化
sudo chown -R postgres:postgres /home/lian/pg
sudo -u postgres mkdir -p /home/lian/pg/pgdata
sudo chown -R postgres:postgres /home/lian/pg/pgdata
@@ -26,36 +30,42 @@ sudo chown -R postgres:postgres /home/lian/pg/pgdata
sudo -u postgres env LD_PRELOAD=/home/lian/try/zvfs/src/libzvfs.so \
/usr/lib/postgresql/12/bin/initdb -D /home/lian/pg/pgdata
## 修改配置文件
cp ./postgresql.conf /home/lian/pg/pgdata/
sudo chown -R postgres:postgres /home/lian/pg/pgdata/postgresql.conf
sudo -u postgres env LD_PRELOAD=/home/lian/try/zvfs/src/libzvfs.so \
/usr/lib/postgresql/12/bin/pg_ctl -D /home/lian/pg/pgdata -l /tmp/pg.log start
sudo -u postgres env LD_PRELOAD=/home/lian/try/zvfs/src/libzvfs.so \
/usr/lib/postgresql/12/bin/psql
sudo -u postgres env LD_PRELOAD=/home/lian/try/zvfs/src/libzvfs.so \
/usr/lib/postgresql/12/bin/pg_ctl -D /home/lian/pg/pgdata -l /tmp/pg.log restart
# 创建测试环境
## 创建测试环境
sudo -u postgres mkdir -p /zvfs/pg_ts_bench
sudo chown -R postgres:postgres /zvfs/pg_ts_bench
sudo chmod 700 /zvfs/pg_ts_bench
CREATE TABLESPACE zvfs_ts LOCATION '/zvfs/pg_ts_bench';
DROP DATABASE IF EXISTS benchdb;
CREATE DATABASE benchdb TABLESPACE zvfs_ts;
DROP TABLE IF EXISTS hook_probe;
CREATE TABLE hook_probe(id int) TABLESPACE zvfs_ts;
INSERT INTO hook_probe VALUES (1);
INSERT INTO hook_probe VALUES (2);
INSERT INTO hook_probe VALUES (3);
INSERT INTO hook_probe VALUES (4);
SELECT * FROM hook_probe;
DELETE FROM hook_probe WHERE id = 1;
UPDATE hook_probe SET id = 11 WHERE id = 2;
SELECT * FROM hook_probe;
## 启动服务
sudo -u postgres env LD_PRELOAD=/home/lian/try/zvfs/src/libzvfs.so \
/usr/lib/postgresql/12/bin/pg_ctl -D /home/lian/pg/pgdata -l /tmp/pg.log start
sudo -u postgres env \
LD_PRELOAD=/home/lian/try/zvfs/src/libzvfs.so \
ZVFS_DEBUG_OPEN=1 \
ZVFS_DEBUG_OPEN_MATCH=PG_VERSION \
ZVFS_DEBUG_RW=1 \
/usr/lib/postgresql/12/bin/pg_ctl -D /home/lian/pg/pgdata -l /tmp/pg.log start
## 打开命令行工具
sudo -u postgres env LD_PRELOAD=/home/lian/try/zvfs/src/libzvfs.so \
/usr/lib/postgresql/12/bin/psql
sudo -u postgres env LD_PRELOAD=/home/lian/try/zvfs/src/libzvfs.so \
ZVFS_DEBUG_OPEN=1 \
ZVFS_DEBUG_OPEN_MATCH=PG_VERSION \
ZVFS_DEBUG_RW=1 \
/usr/lib/postgresql/12/bin/psql
## 重启服务
sudo -u postgres env LD_PRELOAD=/home/lian/try/zvfs/src/libzvfs.so \
/usr/lib/postgresql/12/bin/pg_ctl -D /home/lian/pg/pgdata -l /tmp/pg.log restart
# 5. 验证配置生效
@@ -63,15 +73,76 @@ pid=$(pgrep -u postgres -xo postgres)
echo "pid=$pid"
sudo grep libzvfs /proc/$pid/maps
sudo -u postgres psql -p 5432 -c "show data_directory;"
sudo -u postgres psql -c "SHOW shared_memory_type;"
sudo -u postgres psql -c "SHOW dynamic_shared_memory_type;"
show data_directory;
SHOW shared_memory_type;
SHOW dynamic_shared_memory_type;
```
# 6. 创建测试库(如未创建)
## 简易功能测试
```shell
# 1. 创建 tablespace
DROP TABLESPACE IF EXISTS zvfs_ts;
CREATE TABLESPACE zvfs_ts LOCATION '/zvfs/pg_ts_bench';
sudo -u postgres createdb benchdb
# 2. 创建数据库
DROP DATABASE IF EXISTS benchdb;
CREATE DATABASE benchdb TABLESPACE zvfs_ts;
# 7. 运行你的 bench 脚本
## 3. 连接数据库
\c benchdb
# -- 查看数据库默认 tablespace
SHOW default_tablespace;
# 4. 创建表(默认 tablespace
DROP TABLE IF EXISTS hook_probe;
CREATE TABLE hook_probe(
id int,
val text
);
INSERT INTO hook_probe VALUES (1,'a');
INSERT INTO hook_probe VALUES (2,'b');
INSERT INTO hook_probe VALUES (3,'c');
INSERT INTO hook_probe VALUES (4,'d');
SELECT * FROM hook_probe;
# 5. UPDATE / DELETE 测试
DELETE FROM hook_probe WHERE id = 1;
UPDATE hook_probe SET id = 11 WHERE id = 2;
SELECT * FROM hook_probe;
# 6. 批量 INSERT 测试
INSERT INTO hook_probe
SELECT generate_series(100,10000), 'bulk';
SELECT count(*) FROM hook_probe;
# 7. 创建索引(测试 index file
CREATE INDEX idx_hook_probe_id ON hook_probe(id);
# 8. 大对象测试TOAST
INSERT INTO hook_probe VALUES (99999, repeat('zvfs_test_',1000));
# 9. VACUUM 测试
VACUUM hook_probe;
# 10. ANALYZE
ANALYZE hook_probe;
# 11. 查看物理文件位置
SELECT
relname,
pg_relation_filepath(oid)
FROM pg_class
WHERE relname='hook_probe';
# 12. 删表
DROP TABLE IF EXISTS hook_probe;
# 13. 压测
bash /home/lian/try/zvfs/scripts/run_pgbench_no_mmap.sh
```

47
scripts/do_trace.md Normal file
View File

@@ -0,0 +1,47 @@
```shell
pm=$(pgrep -u postgres -xo postgres)
sudo strace -ff -f -yy -s 256 \
-e trace=%file,read,pread64,readv \
-p "$pm" -o /tmp/pg.sys.trace
strace: Process 233035 attached
strace: Process 235357 attached
^Cstrace: Process 233035 detached
fatal_pid=$(grep 'FATAL: "pg_tblspc' /tmp/pg.log | tail -1 | sed -n 's/.*\[\([0-9]\+\)\].*/\1/p'); echo "fatal_pid=$fatal_pid"
fatal_pid=235357
grep -n "PG_VERSION\|pg_tblspc/16384/PG_12_201909212/16385" /tmp/pg.sys.trace.$fatal_pid
18:access("pg_tblspc/16384/PG_12_201909212/16385", F_OK) = 0
19:openat(AT_FDCWD, "pg_tblspc/16384/PG_12_201909212/16385/PG_VERSION", O_RDONLY) = 4</zvfs/pg_ts_bench/PG_12_201909212/16385/PG_VERSION>
20:read(4</zvfs/pg_ts_bench/PG_12_201909212/16385/PG_VERSION>, "\0\0\0", 4096) = 3
21:read(4</zvfs/pg_ts_bench/PG_12_201909212/16385/PG_VERSION>, "", 4096) = 0
fatal_pid=$(grep 'FATAL: "pg_tblspc' /tmp/pg.log | tail -1 | sed -n 's/.*\[\([0-9]\+\)\].*/\1/p'); echo "fatal_pid=$fatal_pid"
grep -n "\[zvfs\]\[rw-dbg\].*pid=${fatal_pid}" /tmp/pg.log | tail -n 120
6461:[zvfs][rw-dbg][pid=235357][tid=140193133607488][fd=4] read passthrough reason=fd_not_zvfs count=512
6462:[zvfs][rw-dbg][pid=235357][tid=140193133607488][fd=4] read passthrough ret=512 errno=0(OK)
```
```shell
sudo strace -ff -f -k -yy -s 256 -e trace=openat,read,pread64 -p "$pm" -o /tmp/pg.ktrace
```
调用链:
libc(__open64) -> _IO_file_open -> fopen -> AllocateFile -> ValidatePgVersion
是fopen内部调用 libc_(__open64)

View File

@@ -21,7 +21,7 @@ BENCHMARKS="fillrandom,readrandom"
# key数
# NUM=1000000
NUM=500
NUM=5000
# 线程数
THREADS=2

View File

@@ -1,91 +1,53 @@
#!/usr/bin/env bash
set -euo pipefail
# 仅执行 pgbench 的脚本(不安装 PostgreSQL不 initdb不启停服务不改配置
#
# 前提条件:
# 1) PostgreSQL 已经在运行
# 2) 测试库已经存在(默认 benchdb
# 3) PostgreSQL 已经在外部配置为禁用 mmap 共享内存:
# shared_memory_type = sysv
# dynamic_shared_memory_type = sysv
#
# 关于 Malloc0
# - 当前后端是内存虚拟设备,容量有限。
# - 默认参数故意设置得较小,避免一次灌入过多数据。
#
# 关于 LD_PRELOAD
# - USE_LD_PRELOAD_INIT=1初始化阶段pgbench -i启用 LD_PRELOAD
# - USE_LD_PRELOAD_RUN=1 :压测阶段启用 LD_PRELOAD
# - 设为 0 即可关闭对应阶段的 LD_PRELOAD
#
# 用法:
# bash codex/run_pgbench_no_mmap.sh
#
# 可选环境变量(含义):
# PG_HOST=127.0.0.1
# PostgreSQL 服务器地址。
# PG_PORT=5432
# PostgreSQL 服务器端口(默认改为 5432
# PG_DB=benchdb
# 压测数据库名。
# PG_SCALE=2
# pgbench 初始化规模因子(-s越大初始数据越多。
# PG_TIME=20
# 压测持续时间pgbench -T
# PG_CLIENTS=2
# 并发客户端数pgbench -c
# PG_JOBS=2
# 工作线程数pgbench -j
# PG_SUPERUSER=postgres
# 执行 pgbench 的系统用户(通常是 postgres
# LD_PRELOAD_PATH=/home/lian/try/zvfs/src/libzvfs.so
# LD_PRELOAD 目标库路径(你的 zvfs hook so
# PG_BIN_DIR=/usr/lib/postgresql/16/bin
# pgbench 所在目录;不填时自动从 PATH 查找。
# USE_LD_PRELOAD_INIT=1
# 初始化阶段pgbench -i是否启用 LD_PRELOAD1=启用0=关闭。
# USE_LD_PRELOAD_RUN=1
# 压测阶段是否启用 LD_PRELOAD1=启用0=关闭。
# 简化版 pgbench 测试脚本:
# 1) 参数都在本文件顶部配置;
# 2) 直接连接 benchdb
# 3) 执行 pgbench 初始化和压测
PG_HOST="${PG_HOST:-127.0.0.1}"
PG_PORT="${PG_PORT:-5432}"
PG_DB="${PG_DB:-benchdb}"
PG_SCALE="${PG_SCALE:-2}"
PG_TIME="${PG_TIME:-20}"
PG_CLIENTS="${PG_CLIENTS:-2}"
PG_JOBS="${PG_JOBS:-2}"
PG_SUPERUSER="${PG_SUPERUSER:-postgres}"
LD_PRELOAD_PATH="${LD_PRELOAD_PATH:-/home/lian/try/zvfs/src/libzvfs.so}"
PG_BIN_DIR="${PG_BIN_DIR:-$(dirname "$(command -v pgbench 2>/dev/null || true)")}"
USE_LD_PRELOAD_INIT="${USE_LD_PRELOAD_INIT:-1}"
USE_LD_PRELOAD_RUN="${USE_LD_PRELOAD_RUN:-1}"
# ---------------------------
# 固定配置(按需改这里)
# ---------------------------
PG_HOST="127.0.0.1"
PG_PORT="5432"
PG_DB="benchdb"
PG_SCALE="10"
PG_TIME="20"
PG_CLIENTS="2"
PG_SUPERUSER="postgres"
USE_LD_PRELOAD="1"
LD_PRELOAD_PATH="/home/lian/try/zvfs/src/libzvfs.so"
# 可选:优先使用这个目录;为空时自动从 PATH 里找
PG_BIN_DIR="/usr/lib/postgresql/12/bin"
if [[ ! -x "${PG_BIN_DIR}/pgbench" ]]; then
PG_BIN_DIR="$(dirname "$(command -v pgbench 2>/dev/null || true)")"
fi
if [[ -z "${PG_BIN_DIR}" || ! -x "${PG_BIN_DIR}/pgbench" ]]; then
echo "未找到 pgbench设置 PG_BIN_DIR 或把 pgbench 放到 PATH 中。" >&2
echo "未找到 pgbench先安装 PostgreSQL 客户端或修正 PG_BIN_DIR。" >&2
exit 1
fi
run_pgbench_cmd() {
local use_preload="$1"
shift
if [[ "${use_preload}" == "1" ]]; then
run_pg_cmd() {
if [[ "${USE_LD_PRELOAD}" == "1" ]]; then
sudo -u "${PG_SUPERUSER}" env LD_PRELOAD="${LD_PRELOAD_PATH}" "$@"
else
sudo -u "${PG_SUPERUSER}" "$@"
fi
}
echo "当前参数"
echo "当前配置"
echo " host=${PG_HOST} port=${PG_PORT} db=${PG_DB}"
echo " scale=${PG_SCALE} clients=${PG_CLIENTS} jobs=${PG_JOBS} time=${PG_TIME}s"
echo " preload_init=${USE_LD_PRELOAD_INIT} preload_run=${USE_LD_PRELOAD_RUN}"
echo " scale=${PG_SCALE} clients=${PG_CLIENTS} time=${PG_TIME}s preload=${USE_LD_PRELOAD}"
echo
echo "[1/2] 初始化数据(pgbench -i"
run_pgbench_cmd "${USE_LD_PRELOAD_INIT}" \
"${PG_BIN_DIR}/pgbench" -h "${PG_HOST}" -p "${PG_PORT}" -i -s "${PG_SCALE}" "${PG_DB}"
echo "[1/2] pgbench 初始化(-i"
run_pg_cmd "${PG_BIN_DIR}/pgbench" \
-h "${PG_HOST}" -p "${PG_PORT}" -i -s "${PG_SCALE}" "${PG_DB}"
echo "[2/2] 执行压测(pgbench -T"
run_pgbench_cmd "${USE_LD_PRELOAD_RUN}" \
"${PG_BIN_DIR}/pgbench" -h "${PG_HOST}" -p "${PG_PORT}" \
-c "${PG_CLIENTS}" -j "${PG_JOBS}" -T "${PG_TIME}" -P 5 "${PG_DB}"
echo "[2/2] pgbench 压测(-T"
run_pg_cmd "${PG_BIN_DIR}/pgbench" \
-h "${PG_HOST}" -p "${PG_PORT}" -c "${PG_CLIENTS}" -T "${PG_TIME}" "${PG_DB}"

View File

@@ -1,4 +1,8 @@
pgrep -u postgres -x postgres | while read p; do
echo "PID=$p"
sudo grep -m1 libzvfs /proc/$p/maps || echo " (no libzvfs)"
# pgrep -u postgres -x postgres | while read p; do
# echo "PID=$p"
# sudo grep -m1 libzvfs /proc/$p/maps || echo " (no libzvfs)"
# done
for p in 185591 185633 185664; do
echo "== $p"; grep libzvfs /proc/$p/maps || echo "NO_LIB";
done