Get the most out of your Centmin Mod LEMP stack
Become a Member

Sysadmin Do you always reboot after installing kernel updates?

Discussion in 'System Administration' started by deltahf, Oct 23, 2021.

Tags:
  1. deltahf

    deltahf Premium Member Premium Member

    459
    198
    43
    Jun 8, 2014
    Ratings:
    +349
    Local Time:
    10:41 PM
    I know kernel updates are only applied after reboots (unless you're using KernelCare), but it's generally something I don't do unless I'm aware of a specific kernel-related issue that the latest update resolves.

    So, I wanted to ask you guys... do you always reboot to install the latest kernel after you install any kernel updates, or do you just let them get applied occasionally whenever you eventually reboot the server?

     
  2. Mr. Jinx

    Mr. Jinx New Member

    27
    8
    3
    Apr 18, 2021
    Ratings:
    +15
    Local Time:
    4:41 AM
    I always reboot the server after installing updates.
    A reboot on a fast VPS takes a few seconds, and then I'm sure everything installed & loaded correctly.
    No surprises next time your machines has to reboot.
     
  3. eva2000

    eva2000 Administrator Staff Member

    47,498
    10,769
    113
    May 24, 2014
    Brisbane, Australia
    Ratings:
    +16,742
    Local Time:
    1:41 PM
    Nginx 1.19.x
    MariaDB 5.5/10.x
    Really depends on the Kernel update and what it fixes (for non-KernelCare systems that is). KernelCare usually takes care of it though if you pay for a subscription to get rebootless Kernel updates :D
     
  4. Matt

    Matt Well-Known Member

    881
    398
    63
    May 25, 2014
    Rotherham, UK
    Ratings:
    +628
    Local Time:
    3:41 AM
    1.5.15
    MariaDB 10.2
    All my dedicated servers run KernelCare so I don't reboot them.
     
  5. buik

    buik “The best traveler is one without a camera.” Premium Member

    1,452
    394
    83
    Apr 29, 2016
    Flanders
    Ratings:
    +1,235
    Local Time:
    4:41 AM
    Always reboot servers after updates.

    Private and business. Always according to a phased fixed schedule (business). A reboot is often accompanied by maintenance, hardware stress testing, firmware updates and/or hardware replacement where necessary. Total system always redundant. Usually 3+3 clusters on multiple locations. Depends on how critical the service is.

    Not a fan of kernelcare-like solutions. With 0-day code injections with an update window of only 4 hours.

    When a Red Hat kernel is released, it is first tested intensively in a test environment.
    CVE critical based updates are picked up faster. If a producten system does experience an update problem, incidentally anyway, there is sufficient fallback.