PROCEDURE TO PROPERLY SHUTDOWN STORNEXT AND ACCOMPANYING ECOSYSTEM

========================================================================

 

This is a procedure to shutdown all the equipment that is used by the StorNext system.
This StorNext Backup should have been done successfully.

 

1. Ensure you have a successful Stornext backup:

    Run snbackup from command line.

 

2.   Shutdown the secondary MDC first.

     Login into the secondary MDC  through a SSH session, and make sure it is secondary by running the command :
    
    # snhamgr status

    You should get this :
    ========================
    LocalMode=default
    LocalStatus=running
    RemoteMode=default
    RemoteStatus=primary
    ========================

    Now proceed to shutdown the system :

   # shutdown -h now

 

3  Place the Primary MDC in single mode to avoid any HA Failover attempt.

   # snhamgr peerdown
   # snhamgr mode=single

   Verify -->  # snhamgr -m status  -- will give back --   :single:primary:peerdown:unknown:

 

4  Stop the Storage Manager.

   It is better if there is no Library activity, no requests to retrieve/store.  Check this with the commands :

   # fsstate
   # fsqueue -av

   If there is Media in use and/or Storage requests, wait for them to be done or cancel them with the command :

   # fscancel <ReqID>

   Now Stop the Storage Manager with the command :

   # adic_control stop

 

5.  Make sure ALL SAN Clients are OFF.

    Shutdown ALL SAN clients to make sure NO StorNext filesystem is in use.

6.  On the Primary MDC,  stop all Filesystems but the Shared one.

   # cvadmin -e "stop <fsName>"

   Do this for 'snfs1', 'snfs2' and 'snfs3'

   Verify -->  # cvadmin -e select    -- will give you --   :
   =====================================
   List FSS

   File System Services (* indicates service is in control of FS):
   1>*shared-XXXXXX[0]      located on node-1.richards.com:50243 (pid 24892)
   =====================================

 

7.  Shutdown the Primary MDC.

   # sync; sync
   # shutdown -h now

 

8.  Power off the Metadata Array

 

9.  Power off the SAN storage  (Make sure the SAN storage shutdown was done from the MDC).

    Power off the controllers power supplies. Wait until the back-end LEDs are completely out (3 minutes on average).
    This ensures that all cache data has been flushed.
 
 

10.  Power off the tape Library

 

11.  Power off the Fabric Switches

 

12.  Power off the EtherNet Switches.

 

13.  Power off UPS

To bring things up once everything is back connected, the steps are in reverse, from bottom up,
so that each new component that is brought up can be checked, diagnosed and the good state validated,
BEFORE getting the next component up boot.