软件介绍
　　mdstest是软件的元数据操作基准测试工具，用来模拟对文件或者目录的open、stat、close操作，然后报告性能
　　下载软件压缩包：
yum install openmpi openmpi-devel -y
在/root/.bashrc中添加（注意也要添加mdtest的路径）
export PATH=$PATH:/usr/lib64/openmpi/bin/
source /root/.bashrc
下载
[root@lab8105 ~]# wget http://sourceforge.net/projects/mdtest/files/latest/download
解压
[root@lab8105 ~]# tar -xvf mdtest-1.9.3.tgz
修改makefile：
mdtest: mdtest.c
mpicc -Wall -D $(OS) $(LARGE_FILE) $(MDTEST_FLAGS) -g -o mdtest mdtest.c -lm
参数如下：
Usage:  mdtest [-b #] [-B] [-c] [-C] [-d testdir] [-D] [-e] [-E] [-f first] [-F]
[-h] [-i iterations] [-I #] [-l last] [-L] [-n #] [-N #] [-p seconds]
[-r] [-R[#]] [-s #] [-S] [-t] [-T] [-u] [-v] [-V #] [-w #] [-y]
[-z #]
-b: branching factor of hierarchical directory structure
目录树的分支参数
-B: no barriers between phases (create/stat/remove)
不同的阶段没有隔离 (create/stat/remove)；
-c: collective creates: task 0 does all creates and deletes
共同创建: task 0 完成所有的创建和删除工作；
-C: only create files/dirs
只创建文件或目录，不作删除；
-d: the directory in which the tests will run
指出测试运行的目录（若不指定，则默认当前目录）；
-D: perform test on directories only (no files)
只对目录操作进行测试（不包括文件）；
-e: number of bytes to read from each file
从每个文件读出的文件大小
-E: only read files
只读取文件
-f: first number of tasks on which the test will run
首先运行的任务号；
-F: perform test on files only (no directories)
只创建文件，没有目录；
-h: prints help message
输出帮助信息
-i: number of iterations the test will run
测试迭代循环次数；
-I: number of items per tree node
每个树节点包含的项目
-l: last number of tasks on which the test will run
后运行的任务号
-L: files/dirs created only at leaf level
只在目录树的“叶子”层创建文件/目录；
-n: every task will create/stat/remove # files/dirs per tree
每个任务需要再每棵树中create/stat/remove的文件/目录数；
-N: stride # between neighbor tasks for file/dir stat (local=0)
遍历时指定和相邻任务的跨度
-p: pre-iteration delay (in seconds)
每次迭代之间延时（以秒计算）
-r: only remove files/dirs
删除文件/目录
-R: randomly stat files/dirs (optional seed can be provided)
随机遍历文件/目录 ；
-s: stride between the number of tasks for each test
每次测试的任务数的跨度
-S: shared file access (file only， no directories)
共享文件访问（只针对文件操作）；
-t: time unique working directory overhead
记录特定目录的时间开销
-T: only stat files/dirs
-u: unique working directory for each task
为每个任务指定工作目录；
-v: verbosity (each instance of option increments by one)
-V: verbosity value
-w: number of bytes to write to each file
写到每个文件的字节数
-y: sync file after write completion
再写执行完后同步文件到磁盘（同步写）
-z: depth of hierarchical directory structure
目录树的深度；
NOTES:
* -N allows a "read-your-neighbor" approach by setting stride to
tasks-per-node. Do not use it with -B， as it creates race conditions.
允许"read-your-neighbor" 方法
* -d allows multiple paths for the form '-d fullpath1@fullpath2@fullpath3'
可以指定多个测试路径，'-d fullpath1@fullpath2@fullpath3'
* -B allows each task to time itself. The aggregate results reflect this
change.
允许每个任务对自己进行计时；
* -n and -I cannot be used together.  -I specifies the number of files/dirs
created per tree node， whereas the -n specifies the total number of
files/dirs created over an entire tree.  When using -n， integer division is
used to determine the number of files/dirs per tree node.  (E.g. if -n is
10 and there are 4 tree nodes (z=1 and b=3)， there will be 2 files/dirs per
tree node.)
* -R and -T can be used separately.  -R merely indicates that if files/dirs
are going to be stat'ed， then they will be stat'ed randomly.
不能同时使用，因为指定了每个树节点的文件/目录数量，而-I指定的是整棵树的文件/目录数量。
Illustration of terminology:
Hierarchical directory structure (tree)
=======
|       |  (tree node)
=======
/   |  
------    |    ------
/          |         
=======     =======     =======
|       |   |       |   |       |    (leaf level)
=======     =======     =======
In this example， the tree has a depth of one (z=1) and branching factor of
three (b=3).  The node at the top of the tree is the root node.  The level
of nodes furthest from the root is the leaf level.  All trees created by
mdtest are balanced.
　　这个例子中，目录树深度为1（z=1），每个节点分支为3（b=3）.上面的节点为根节点。