Forum Discussion

vj_singh_177738's avatar
vj_singh_177738
Icon for Nimbostratus rankNimbostratus
Jan 02, 2015

what is the best method to reboot F5 LTm

I would like to know what is the best process to reboot F5 LTM box.

 

Do I need to halt the services first and then reboot. or directly reboot. it.

 

10 Replies

  • nathe's avatar
    nathe
    Icon for Cirrocumulus rankCirrocumulus
    Jan 02, 2015

    There's a few ways.

     

    GUI - System - Configuration - Reboot.

     

    Else, from the command line I would either use the reboot command or "shutdown -h now" (this will halt and shutdown the BIG-IP). So, to clarify I don't believe you need to Halt the system if rebooting via the command line or GUI.

     

    Finally there is the LCD display - here you would Halt the system first before powering off or rebooting.

     

    Hope this helps, N

     

  • hari_126827's avatar
    hari_126827
    Icon for Cirrus rankCirrus
    Jan 02, 2015

    Hi,

     

    The best way to reboot F5 LTM (individual or in clusters) is to:

     

    • first take archive (via F5 GUI: System > archives) and ltm or tmm errors in logs and make sure to have console connection in case of nonreachability..
    • and then login to F5 LTM via ssh and then go to bash shell (by typing command 'run util bash' command)
    • take the output of existing running services with "show sys service | grep run" and then type 'reboot' command.
    • execute command "show sys service | grep run" and compare results with above step, in case of any deviations check with solution given to restart the services with 'bigstart' utility command

    Alternative is to do it from GUI but is slow compared to bash shell mode., Please check the link for further info: SOL7369: Shutting down and restarting the BIG-IP system

     

    SOL4080: Adding and removing services included in the 'bigstart' process

     

  • StephanManthey's avatar
    StephanManthey
    Icon for Nacreous rankNacreous
    Jan 02, 2015

    Before rebooting your BIG-IP please make sure you can control the boot process from CLI by attaching a serial console (never ever change it from the defaults 19.200 8N1 settings) allowing you to select the boot volume of your choice.

    In case you do not have a serial console you may consider to monitor the boot process via SSH connection to the AOM (Always On Module).

    For unattended reboots please make sure in advance you are booting into the right boot volume by running: 

    switchboot -l

    It will list the available and default boot volume of your system.

    For a pure reboot the following command from CLI will do the job: 

    shutdown -r now

    If you use the "-h" switch, the unit will halt and you need to power it on via AOM/SCCP (serial or SSH connection required) or by pushing a button on the appliance front panel (might be a challenge with device in remote location!).

    A reboot may be followed by file system checks (and depending on upgrade / downgrade operations changes to BIOS, MOS, build-in FPGA followed by additional automated reboots.)

    Be very careful if your command prompt contains exclamation marks ("!"). Upgrades to components may still be in progress.

    After a reboot I used to monitor the /var/log/ltm ("tail -f /var/log/ltm") to check everything is finished before applying new changes to the system.

    • StephanManthey's avatar
      StephanManthey
      Icon for Nacreous rankNacreous
      Jan 02, 2015
      Your unit will be completely unavailable during the reboot cycle! And depending on hardware, software, configuration elements and required component updates it will take at least a couple of minutes to get it back in service. If your are running your system in production, make sure you have initiated a failover to the peer unit of your sync-failover device-group (assuming you are on v11) and the other unit is active and handles the traffic. Without redundancy your service(s) provided via BIG-IP will be completely offline during the reboot cycle. If you do a failover, please verify your peripheral components (routers, firewalls, servers) get their ARP cache updated (may change without mac-masquerading if a virtual address [for VIP, SNAT, NAT, floating Self IP] moves to the failover unit and the related gratuitous ARPs from the F5 device get lost for some reason).
  • Bret_Linne_3433's avatar
    Bret_Linne_3433
    Icon for Nimbostratus rankNimbostratus
    Mar 01, 2018

    Is there a way to reboot the BigIP through an ansible Playbook? This is in the case of installing a license remotely from an ansible controller, then sending a command to the BigIP to reboot it to reload the license properly. How might this be done?

     

    • StephanManthey's avatar
      StephanManthey
      Icon for Nacreous rankNacreous
      Mar 02, 2018

      Hi Bret, you may want to use the command "

      reloadlic
      " to reload a new license file. Perhaps the provision of new features/modules may require a "
      reboot
      " or "
      shutdown -r now
      ". Cheers, Stephan

  • JG's avatar
    JG
    Icon for Cumulonimbus rankCumulonimbus
    Mar 01, 2018

    For years I have been running the command recommended by F5: "full_box_reboot".

    I have just found that 

    /usr/bin/full_box_reboot
    is just a shell script containing a single command: "reboot". 🙂 )