I have 5 different VPS accounts that I manage with Vultr, and 1 with Digital Ocean.
I'm using the single core 512mb of RAM w 300+ mb of swap in Vultr.
DigitalOcean specs are 2 core, 2 GB.
Ubuntu 16.04
crontab -e is set to reboot in root every day at 3 (I don't know if that's EST)
The websites all randomly go down, but only at 2:23-2:26 am and they all come back online 11pm the following day (I'm assuming that's my reboot, not sure).
I can't find any source for the crashes. Error logs look ok to me, but it's been making me pull my hair out for a few months now. The sites get little to no traffic, except for bots possibly
What can cause this? How can I get the problem resolved? Has anyone heard of something similar or know where to point me to at least start looking?
11 websites. 5 servers, all independent of each other, all randomly doing the exact same thing.
Server is completely offline, not just httpd but ssh, etc.
Thank you to whoever helps with this. I may have a few extra bucks in my PayPal () to whoever can stop this.
Someone smarter than me save me from what is probably obvious.
What is in common to all? Any firewall bans being shared? Could anything be updating DNS incorrectly? DNS can take about a day or more to timeout sometimes. But if that was the case, you'd still be able to ssh/ping the IP itself. Does pinging domain result in correct IP when its down? When down, are any of the IPs ping-able?
The crontab -e is going to take effect at 3.am local time on the server, depending on its location.
512mb isn't going to be enough if you're running MySQL. I never even tried 512mb RAM myself. I tried 1GB a few times, but it wasn't stable. Websites didn't crash, but Virtualmin would go down every other day.
Your syslog should have entries related to this.
I provide FREE Server Management Services in exchange of money
I understand 512 mb may not be enough, but that doesn't explain why every site on a few completely separate servers goes down simultaneously, despite having zero links to each other.
I was hoping someone was going to tell me that there's a setting or something that happens at 2:27 AM EST that Virtualmin DOES overflow the RAM.
Weirdly, just one of my servers doesn't do it. 3 sites on that one specifically too, one with more mySQL use than average.
Fail2ban is disabled and I don't know what you mean by the DNS causing a server crash.
If I ping it doesn't do anything. Entire server down even ssh etc.
Ty for your help