Skip to main content

Uptime vs Availability

Do we need to sacrifice system uptime against system availability. And I use the term system availability to identify both uptime and unscheduled outages. So here system availability includes time from both unscheduled and scheduled outages. Why would I ask such a thing? Systems crash for all kinds of reasons: failure in the garbage collector to collect objects or collecting the wrong objects. So why do we need to do regular scheduled reboots of the system?

How about to ensure that the system is configured correctly to handle a reboot automatically. Frequent patches, installation of new software and configuration changes can all affect the boot process. And if the system hasn't been booted in the last quarter, the last 90 days, do you know that it will successfully boot and launch the operating system and all services? What about adding and removing (and even changing the configuration) of services and applciations? Will taht cause problmes? For linux, changes in mount-points could cause additional problems... as is changes in the remote systems a server depends on? Does taking a premptive boot monthly or quarterly reduce your risk and guarantee more system availabilty at the expense of uptime? Is it worth it? Better yet, can you take the risk and have a significant outage against your overall system avaialability for the sake of uptime?