Dealing With Server's Power Outages Is Kinda... Stressful
About week ago i encountered a critical problem in my work. The LED indicator has several amber lights on. I proceed to boot the server up while LED lights are still blinking. My monitor doesn't even show anything even POST message won't showed off. I began to open up server hood and turns out the sysboard aka mainboard are really out of business.
Night before, I was trying to work from home by using VPN and SSH from my macbook but the connection isn't settled. I'm suspiciously thinking that power outages are happened already. Though server was backing up with UPS device, it doesn't really helpful if the outages duration are long enough. Then the nightmare happened, my domain controller server is down. Hell broke loose.
I need two days to put everything back to normal. I let every computer to be run using domain still, but changing hosts configuration in every computer is a must, or else every user cannot accessing our internal ERP site which is gonna be the big problem. This article's title are kinda exaggerate, but it actually stressful. Fortunately i made it, though.
Till this day i still figure out how can i prevent these power outages gonna happen again. Doing some research and these pro's are recommending to simply turning on a generator, which is quite impossible to do since there are nobody in my office would stand by and make it up at night. I come out with some solution that "might" work, by using wake on LAN feature, which is using main router to send a "magic package" via ethernet and the signal are gonna translated as "turn this server on" to my server through local network. Next step is schedule it to be activated at some time, which in my case are at 7 AM. The server itself are gonna be running only through work hour, and i already set it to turn off at 6 PM by using "crontab" in my Linux Debian 9.
This is the best-temporary fix, maybe until i propose bigger UPS loads.