On Fri, 2011-03-11 at 21:25 +0100, Per Jessen wrote:
I surmise that you have probably never worked in an environment with such SLAs. Boot time is rarely if ever a factor in the evaluation of risk. The only/best way to keep uptimes of 99.998% or more is 1) to reduce change to an absolute minimum and 2) when at all necessary, make sure changes happen during the planned 30min outage window on Christmas Eve.
Wrong, IMNSHO. It is the service that is supposed to have 99.998% availability, not the machine. Anyone not using some sort of HA solution, with redundant machines, is battling statistics, and will eventually lose. For this reason, you are free to run updates with significant downtime on passive machines, and then fail them over, until all machines have been updated. I know sysadmins who run things the way you describe, running updates directly on the live system, causing completely unnecessary outages. If the service really is necessary, that is completely the wrong way of doing it Anders -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org