Nginx Site is down, CPU High

HTOP: https://instashot.io/i/24w48.png

---

boot.log errors:

Code:

No kdump initial ramdisk found.[WARNING]

Rebuilding /boot/initrd-2.6.32-573.26.1.el6.x86_64kdump.img
mktemp: failed to create file via template `/tmp/initrd.img.XXXXXX': No space left on device
mkdir: cannot create directory `/tmp/initrd.u9PWbm/var': No space left on device
Saving random seed failed.
Failed to run mkdumprd

df -h

Code:

Filesystem            Size  Used Avail Use% Mounted on
/dev/vda1              99G   15G   79G  16% /
tmpfs                 3.8G     0  3.8G   0% /dev/shm
/home/usertmp_donotdelete
                      3.7G 1003M  2.6G  28% /tmp

No idea what's going on. I was under an attack. I enabled CloudFlare attack mode and it seems to be working, however the server is still affected by the attack even though there are no attack hitting it now.

I tried to do ls in /tmp and it wouldn't even show me what's there. It must be full. Can I run rm -rf on tmp then reboot with no issues?

 

Deleted tmp, and it seems to bring the site online. Any downsides doing so?

 

I ran rm -rf * in there and even though it gave that error i assume it deleted some? i also ran other commands to try to delete it such as find.

once i did that it came back online quickly, it seems like mysql was causing high cpu usage as well as php since the directory was full perhaps? it had thousands and thousands, maybe even milions of files.

i was under attack, called "PHP Session Attack" it generate tons of unused sessions.