I noticed today when running uptime that my server has been up for exactly 365 days. Which led me to google "should I periodically reboot my server". Which led me to the knowledge that YES, I should reboot after any kernel update, or the new kernel doesn't become active. Kinda obvious I guess, but not something I knew before today. I've always kept on top of my updates, but never reboot...

...and now I'm a bit scared to.

I rebooted a different box a while ago, and it failed to come back up due to some grub conf problem. After failing to fix it from rescue mode (I had no idea what I was doing) I opted just to reinstall the box - there wasn't much on it anyway and it was trivial to do so.

Is there anything I should check (in config files etc) before I reboot that might minimise my chances of a similar failure?


Typically you want to have an out-of-band management session that will allow you to monitor the reboot. Ensure that you have the root password at hand in case you do need to enter single user mode.

You might want to check if the reboot will force a file-system check, which might take a long time on very large file-systems.

Similar, you might want to check for the presence of a file /.autorelabel which will trigger a potentially lengthy SELinux relabeling of the file-systems. So will a change from having disabled SELinux to re-enabling it. A modified /etc/selinux/config not matching the status given getenforce shows might potentially be slightly disruptive.

And you might want to confirm that services currently running (or not) match those found in chkconfig --list

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.