During our meeting today we noticed that Videl appeared to be down. All ports appeared to be closed, but ICMP ping packets were responding.
Strangely though, when I got home Videl appeared to be working 100% again. After checking the logs, I realized the Videl downtime was due to a mistake I made 2 months ago! I had made a vserver enhanced kernel in order to run virtual server partitions on Videl, separating mail, web and database servers into their own security isolated partitions. The theory is that even if an individual server is cracked, they cannot do anything to other vserver partitions on the same box. 2 months ago I must have manually booted into this vserver kernel, found it to be stable and went home, forgetting to make it the default kernel. The power must have gone out for an extended period of time today. When the server came back up, it booted the default Red Hat kernel without vserver capability, meaning none of the web, mail or database servers came up either. Ray noticed this and quickly fixed it. Thanks Ray! Moral behind this story? Yeah, it is unlikely that your Linux server will ever need rebooting, but don't depend entirely on high uptime to save you. Be sure to check that your server will reboot properly. Warren Togami [EMAIL PROTECTED]
