Cray System Management Documentation > Cray System Management (CSM) Administration Guide > power management > Power Off the External Lustre File System

Power Off the External Lustre File System

General procedure for powering off an external ClusterStor system.

Use this procedure as a general guide to power off an external ClusterStor system. Refer to the detailed procedures in the appropriate ClusterStor administration guide:

Title	Model
`ClusterStor E1000 Administration Guide 4.2 - S-2758`	ClusterStor E1000
`ClusterStor Administration Guide 3.4 - S-2756`	ClusterStor L300/L300N
`ClusterStor Administration Guide - S-2755`	Legacy ClusterStor

Procedure

(remote#) SSH to the primary MGMT node as admin.
```
ssh -l admin cls01234n00.us.cray.com
```
(n000$) Change to root user.
```
sudo su –
```
(n000#) Collect status information for the system before shutdown.
```
cscli csinfo
cscli show_nodes
cscli fs_info
crm_mon -1r
```

(n000#) Check resources before unmounting the file system.

ssh cls01234n002 crm_mon -r1 | grep fsys
ssh cls01234n004 crm_mon -r1 | grep fsys
ssh cls01234n006 crm_mon -r1 | grep fsys
ssh cls01234n008 crm_mon -r1 | grep fsys
ssh cls01234n010 crm_mon -r1 | grep fsys
ssh cls01234n012 crm_mon -r1 | grep fsys
. . .

(n000#) Stop the Lustre file system (FILESYSTEM_NAME will be reported from the cscli fs_info command run above).
```
cscli unmount -f FILESYSTEM_NAME
```

(n000#) Verify that resources have been stopped by running the following on all even-numbered nodes.

ssh NODENAME crm_mon -r1 | grep fsys

Example output:

cls01234n006_md0-fsys (ocf::heartbeat:XYMNTR): Stopped
cls01234n006_md1-fsys (ocf::heartbeat:XYMNTR): Stopped
cls01234n006_md2-fsys (ocf::heartbeat:XYMNTR): Stopped
cls01234n006_md3-fsys (ocf::heartbeat:XYMNTR): Stopped
cls01234n006_md4-fsys (ocf::heartbeat:XYMNTR): Stopped
cls01234n006_md5-fsys (ocf::heartbeat:XYMNTR): Stopped
cls01234n006_md6-fsys (ocf::heartbeat:XYMNTR): Stopped
cls01234n006_md7-fsys (ocf::heartbeat:XYMNTR): Stopped

(n000#) SSH to the MGS node (the MGS_NODE name will be reported from the cscli fs_info command run above).
```
ssh MGS_NODE
```
(mgs#) To determine if Resource Group md65-group is stopped, use the crm_mon utility to monitor the status of the MGS and MDS nodes.
1. Shows MGS and MDS nodes in a partial stopped state.
```
[MGS]# crm_mon -1r | grep fsys
```
  Example output:
```
cls01234n003_md66-fsys (ocf::heartbeat:XYMNTR): Stopped
cls01234n003_md65-fsys (ocf::heartbeat:XYMNTR): Started
```
2. If the output of the previous command shows a partial stopped state (Stopped and Started), issue the stop_xyraid command and verify that the node is stopped.
```
[MGS]# stop_xyraid nodename_md65-group
[MGS]# crm\_mon -1r | grep fsys
```
  Example output:
```
cls01234n003_md66-fsys (ocf::heartbeat:XYMNTR): Stopped
cls01234n003_md65-fsys (ocf::heartbeat:XYMNTR): Stopped
```
(mgs#) Exit the MGS node.
```
exit
```
(n000#) Power off the non-MGMT diskless nodes.
1. Check power state of all non-MGMT nodes and list the node hostnames (in this example cls01234n[02-15]) before power off.
```
pm -q
```
  Example output:
```
on: cls01234n[000-001]
on: cls01234n[002-015]
unknown:
```
2. Power off all non-MGMT nodes.
```
cscli power_manage -n cls01234n[02-15] --power-off
```
3. Check the power status of the nodes.
```
pm -q
```
  Example output:
```
on: cls01234n[000-001]
off: cls01234n[002-015]
unknown:
```
(n000#) Repeat the previous step until all non-MGMT nodes are powered off.
(n000#) From the primary MGMT node, power off the MGMT nodes.
```
cscli power_manage -n cls01234n[000-001] --power-off
```
(n000#) Shut down the primary management node.
```
shutdown -h now
```

Next step

Return to System Power Off Procedures and continue with next step.