Single server uptime is a poor way to determine the health of a production environment. It is common thought that production environments should be built to fail. This is where multiple servers and load balancing come into play. The real measurement of uptime is how long the production environment is up. In many cases you could replace servers, networks, and any other component. As long as the customer/end-user does not know, then you can measure it as success. I understand you only have the one server. I know most of us have been there. We are saying that there is a better way. However, I don't know your circumstances. So we will try to clarify the alternatives with a singles server. As I see it these are your options. First, schedule downtime, and reboot the whole server. Second, reboot services manually, log out of all users, and finally use init 3 to kill the grapical interfaces. Not all three steps are always needed, but I'm not going to deep dive into when to do each here. Finally, suse live patching. The last option is currently only available in SLES12, unless someone can get kgraft working on opensuse. That would be awesome :-) -----Original Message-----From: jdd <jdd@dodin.org> To: opensuse@opensuse.org Subject: Re: [opensuse] How to quickly and easily restart all the demons and processes that got affected bey zypper up? Date: Tue, 23 Jun 2015 18:37:09 +0200 Le 23/06/2015 18:29, Per Jessen a écrit :
jdd wrote:
I always wonder how people can have uptime more than one year when so many kernel updates run around :-(
Redundant servers with load balancing. It's not black magic.
how does this makes uptime better? jdd -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org