多节点计算提示 Your Open MPI job may now hang or fail

naxiangzi

centos8.5 + slurm24 +openmpi5

想看一下配置， 也没有mpd.conf, openmpi5 是编译安装，没发现错误
[root@master~]# find / -name mpd.conf
find: ‘/run/user/1000/gvfs’: 权限不够

1. chkpt1
   
2. chkpt2
   
3. chkpt3
   
4. chkpt4
   
5. --------------------------------------------------------------------------
   
6. WARNING: Open MPI failed to TCP connect to a peer MPI process.  This
   
7. should not happen.
   
8. 
   
9. Your Open MPI job may now hang or fail.
   
10. 
    
11.   Local host: master
    
12.   PID:        268572
    
13.   Message:    connect() to 192.168.2.101:1073 failed
    
14.   Error:      Connection timed out (110)
    
15. --------------------------------------------------------------------------
    

复制代码