One #Kernel to rule them all …
#software #difference #microsoft #windows #freedom #foss #linux #os #software #knowledge #meme
One #Kernel to rule them all …
#software #difference #microsoft #windows #freedom #foss #linux #os #software #knowledge #meme
I was a Linux admin on a team of fourteen. A Windows Server admin bragged at his server had gone a full two months without a reboot.
In fact, it was 65 days. The other Windows admins congratulated him. He came over to tell me the news, which I already overheard and sshed in to check my uptime.
“821 days uptime.” He got disappointed. That’s cool, because he was a jerk for many reasons.
But deep inside, I realized I needed to stay late after the call center closed, as it is a good idea to apply kernal patches to your firewall.
821 days is a stress inducing amount of uptime
If you want high availability build a cluster and then update and reboot individual machines
It wasn’t that bad. And I had two routers for failover. That was the failover router; I just wanted to be an asshole towards the douchebag who has a wife at home with a newborn baby, who hit on women 15 years younger than him constantly.
So I b-lined to the machine that I knew would have the longest uptime.
That’s more than 2 years. Maybe I’m just being hash but surely you would reboot the device at least once a year
In a production environment 20 years ago, without a clustered environment? Hell no.
Hell, I have an airgapped system at home I haven’t restarted in a decade.
That’s really bad for all sorts of reasons but I’m sure you probably know that now
Yeah I was going to say, even just as a homelab admin, I can keep my OS uptime high if I want to, and just update my services, but I’ve missed some auto start fixes etc., where I updated and started the updated service manually, but then next power loss it didn’t start properly.
Now I try to make sure I reboot once a month when it’s convenient for me to fix issues to make sure all updates are applied and tested. And of course kernel updates like you say.