Anyways, up until yesterday, we were running the Astaro v8 beta software, but we decided to upgrade all the way up to 8.001. Today, I needed to do a port scan on the internal network while I was outside of it, so I fired up the VPN, and ran the following nmap:
> nmap -sS -p1-65535 192.168.0.0/24
To my surprise, about halfway through my NMAP I got a notice that my connection to the VPN had been dropped. I assumed that either I had triggered some security feature or this had been a coincidence, so I asked a coworker to attempt to VPN to the machine; no connection. About this time Nagios (sitting outside the VPN) started freaking out about all of the machines inside the VPN (NRPE - using simple port forwarding). I went to vSphere and hit Astaro's console only to find that it wouldn't respond to any key presses at all (it was at the Astaro splash screen). I finally had to reboot it in vSphere.
I believed, at the time, that the problem was just coincidental to my scan, after all we had done the same exact scan many times in the past, but a second attempt caused the same behavior. On a third attempt, I tail -f'd all of my log files and, other than a rapidly exploding packet filter log, nothing unusual was reported up to the very moment that everything just froze up.
I used vSphere's Performance tab and noticed a few things. First, around the time my nmap started, the CPU usage on Astaro jumped to 100% and stayed there even after the machine froze. Now, the server that this is on has two 2ghz quadcore CPUs and Astaro has full access to all of those resources as it needs it, so I doubt the box's performance is an issue. Disk space is not large at all (~62% of the allocated space).
I'm not ready to rule out VMWare as the culprit here, but I do want to remind anyone reading this that we never had this problem with the v8 beta versions even when we were running much higher network traffic.
