1

bin/run-fs-benchmark.sh

@@ -264,6 +264,7 @@ export LIBHDFS_OPTS="${final_java_opt}"
 # see https://github.com/apache/doris/blob/master/docs/zh-CN/community/developer-guide/debug-tool.md#jemalloc-heap-profile
 export JEMALLOC_CONF="percpu_arena:percpu,background_thread:true,metadata_thp:auto,muzzy_decay_ms:30000,dirty_decay_ms:30000,oversize_threshold:0,lg_tcache_max:16,prof_prefix:jeprof.out"
 export AWS_EC2_METADATA_DISABLED=true
+export AWS_MAX_ATTEMPTS=2
 
 echo "$@"
 ${LIMIT:+${LIMIT}} "${DORIS_HOME}/lib/fs_benchmark_tool" "$@" 2>&1 | tee "${LOG_DIR}/fs_benchmark_tool.log"

docs/en/docs/lakehouse/fs_benchmark_tool.md

@@ -84,27 +84,29 @@ The type of file system on which the operation is required. Currently supported
 
 Specifies the type of operation
 
--  `create_write`: Each thread creates a file named `test_${current thread number}`  in the `base_dir(set in conf file)` directory and writes to the file with a write size `file_size` of.
+- `create_write`: Each thread creates a file named `test_${current thread number}`  in the `base_dir(set in conf file)` directory and writes to the file with a write size `file_size` of.
 
--  `open_read`: On `create_write` the basis of the created file, each thread reads the file with the name of `test_${current thread number}`  and the read size of `file_size`.
+- `open_read`: On `create_write` the basis of the created file, each thread reads the file with the name of `test_${current thread number}`  and the read size of `file_size`.
 
--  `exists`: Each thread queries whether a file with  `test_${current thread number}` filename exists.
+- `single_read`: Read `file_path(set in conf file)` file, read size is `file_size`.
 
--  `rename`: On `create_write` the basis of the created file, each thread changes the `test_${current thread number}` filename to `test_${current thread number}_new`.
+- `prefetch_read`：Use prefetch reader to read `file_path(set in conf file)` file, read size is `file_size`. Only for s3 file system.
 
--  `single_read`: Read `file_path(set in conf file)` file, read size is `file_size`.
+- `exists`: Each thread queries whether a file with  `test_${current thread number}` filename exists.
 
--  `list`: Get `base_dir(set in conf file)` the list of files in the directory.
+- `rename`: On `create_write` the basis of the created file, each thread changes the `test_${current thread number}` filename to `test_${current thread number}_new`.
+
+- `list`: Get `base_dir(set in conf file)` the list of files in the directory.
 
 `--file_size` 
 
 The file size of the operation, in bytes.
 
--  `create_write`: Default is 10 MB.
+- `create_write`: Default is 10 MB.
 
--  `open_read`: Default is 10 MB.
+- `open_read`: Default is 10 MB.
 
--  `single_read`: The default is 0, that is, the full file is read.
+- `single_read`: The default is 0, that is, the full file is read.
 
 `--threads`
 

docs/zh-CN/docs/lakehouse/fs_benchmark_tool.md

@@ -85,12 +85,14 @@ sh run-fs-benchmark.sh \
 
 - `open_read`：在`create_write`创建好文件的基础下，每个线程读取文件名为`test_当前的线程号`的文件，读取大小为`file_size`。
 
+- `single_read`：读取`file_path(conf文件中设置)`文件，读取大小为`file_size`。
+
+- `prefetch_read`：使用 prefetch reader 读取`file_path(conf文件中设置)`文件，读取大小为`file_size`。仅适用于 s3。
+
 - `exists` ：每个线程查询文件名为`test_当前的线程号`的文件是否存在。
 
 - `rename` ：在`create_write`创建好文件的基础下，每个线程将文件名为为`test_当前的线程号`的文件更改为为`test_当前的线程号_new`。
 
-- `single_read`：读取`file_path(conf文件中设置)`文件，读取大小为`file_size`。
-
 - `list`：获取 `base_dir(conf文件中设置)` 目录下的文件列表。
 
 `--file_size`