# LSF

## 1.命令行提交LSF任务

* `-J`LSF任务名称
* `-M`该任务最大内存限制数，最大不能够超过队列限制，队列限制内存可使用`bquences -l quence name`查看
* -e 错误日志文件
* -o 错误输出文件
* `-R span[hosts=1]` 单个节点进行计算，一般不会涉及到跨节点
* `-q quence name`指定队列
* `-n`

```bash
bsub 
-J extract
-n 1
-M 200GB
-e test.err
-o test.out 
-R span[hosts=1]
-q high 
"python extract_count2.py 111111.gz 2222.gz" //双引号可省略
```

## 2.文件提交LSF任务

```bash
#BSUB -J R1extractCount
#BSUB -n 1
#BSUB -R span[hosts=1]
#BSUB -o extract.out
#BSUB -e extract.err
#BSUB -q "smp"
python extract_count2.py 111111.gz 2222.gz //要运行的脚本
```

## 3.大批量提交任务

* 先给LSF提交一个for循环，告诉它有很多任务需要提交，让它自动根据系统的资源使用情况来对任务进行分配；从而避免多个任务同时记在一个节点，造成系统负荷

  ```bash
  #BSUB -J R1extractCount
  #BSUB -n 1
  #BSUB -R span[hosts=1]
  #BSUB -o extract.out
  #BSUB -e extract.err
  #BSUB -q "smp"
  A2files=`ls .|grep "count.txt"`
  for i in ${A2files[@]};do
  {
   sh ./binomTest.sh ${i}
   sleep 5
  }&   //这里并行同时提交                                           
  done
  ```
* for循环中，每个item都运行一个脚本

  `binomTest.sh`进行任务的提交

  ```bash
  //${1} 是该脚本接受的第一个参数
  bsub -K -J binomalTest_${1} -n 20 -e binomalTest_${1}.err -o binomalTest_${1}.out -R span[hosts=1] -q "high" "python test.py"
  ```

## 4.常用命令

1. 检查提交作业状态

   `bjobs`
2. 显示该作业的所有信息

   `bjobs -l jobsID`
3. 删除不需要的作业

   `bjobs kill jobID`
4. 监视作业运行

   `bpeek jobsID`
5. 显示队列信息

   `bqueues`

   `bqueues –l <queue name>`
6. 显示各节点作业相关情况

   `bhosts`

   `bhosts <hostname>`
7. 显示各节点负载信息

   `lsload`

   `lsload <hostname>`

## 参考

1. <https://blog.csdn.net/cy413026/article/details/84649514>&#x20;
2. [https://www.ibm.com/developerworks/cn/linux/l-lo-efficient-cluster-manage-system-LSF/index.html#1%E7%AE%80%E5%8D%95%E7%9A%84%E4%BD%BF%E7%94%A8outline](https://www.ibm.com/developerworks/cn/linux/l-lo-efficient-cluster-manage-system-LSF/index.html#1简单的使用outline)&#x20;


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://zpliu.gitbook.io/booknote/linux/lsf-ren-wu-ti-jiao-xi-tong.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.