Restarting an unresponsive server

NAS Platform and Unified Storage Series 4000 Hardware Reference

Version
14.6.x
Audience
anonymous
Part Number
MK-92HNAS030-27

Perform this process to restart an unresponsive server from the server operating system (OS) console. You generate a diagnostic log that can help you better understand the problems. You can gain access either by using SSH software to connect to the server's CLI or connecting to the server serial port.

  1. Connect to the SMU using the ssh software.
  2. From the siconsole, select the server.
    • If the system fails to respond, go to step 3.
    • If the system takes you to the server OS console, issue the command: bt active, so you can view the display.
    • If you are still at the siconsole, select q, press Return, and then perform the following steps:
      1. Connect directly to the MMB as manager using ssh.
      2. If the connection succeeds, you are taken to the server OS console, where you issue the command: bt active
      3. If the connection fails, continue to step 4.
  3. Connect to the system with a serial null modem cable, and perform the following steps:
    See Serial port if you need details.
    1. Login as manager or you will get the Linux prompt, not the server OS.

      If you use root, use ssc localhost.

    2. Issue the command: bt active
  4. If you are still unable to get to the server OS, perform the following steps:
    1. Check to make sure that the Bali CLI is booting successfully.
    2. Log in through the serial cable connection.
    3. Tail /var/opt/mercury-main/logs/dblog
    4. Search the log for the entry MFB.ini not found run nas-preconfig.
      • If the entry is present, the system has been unconfigured by either running the unconfig script or removing the node from a cluster.
      • If the entry is not present, monitor the dblog during the boot cycle to see where it fails.
    Warning: If the server is still unresponsive, do not pull the plug. Instead, see the next step. The reboot time varies from system to system. The reboot can take up to 20 minutes, because a dump is compiled during the reset process.
  5. Check the green LED on the front panel of the server for the server status.
    Table. Server status - model HNAS 4040
    Server status LED Meaning
    Amber Critical failure and server is not operational.
    Slow green flash (every three seconds) System has been shut down and you can remove the power.
    Medium green flash (every .8 seconds) Server needs attention, and a non-critical failure has been detected. For example, a fan or power supply has failed.
    Off Normal operation.
    Table. Server status - models HNAS 4060, 4080, and 4100
    Server status LED Meaning
    Solid green System has rebooted and the server is online.
    Slow green flash (every three seconds) System has been shut down and you can remove the power.
    Medium green flash (every 1.3 seconds, with the server online) No enterprise virtual server (EVS) residing on the server.
    Medium green flash (every .8 seconds) Server is available.
    Fast green flash (5 times per second) Server is booting.
  6. If the green LED is flashing 5 times per second, plug in the serial cable.
    • If the terminal screen is generating output, let the process complete.
    • If the terminal screen is blank, press the Reset button.
      Note: Pulling the power cord from the server is not recommended. Do not pull the power cord unless it is absolutely necessary. First, complete the steps above.