Hey everyone, I recently experienced a significant downtime on my mail server, which I've been managing as a student for years. It runs Postfix for SMTP and Dovecot for IMAP, and I keep a close watch on it. Typically, I've only dealt with minor issues like updates gone wrong or disk space problems, but this incident lasted a whopping 31 hours! This happened while I was on a business trip, so I missed out on my usual email checks and my monitoring didn't catch any downtime (as Postfix was still responding to connections, even though it had TLS issues). I found a warning in the logs saying "no entropy for TLS key generation", which stopped email from flowing in or out. Since my server is a Cloud VM that's been reliable until now, I'm curious about why it suddenly ran out of entropy and how I can prevent this from happening again. Any advice on making sure it fails more gracefully in such cases would be appreciated!
1 Answer
Sounds like you're on Ubuntu; users often get warnings about low entropy. You might want to install 'haveged' to help boost your entropy pool. Keeping an eye on that would definitely help! If you can, look into physical random number generators, but I know that can be costly depending on your setup.
Thanks! I'll definitely check out 'haveged'. I'm trying to keep costs down, so a physical machine isn't in the cards right now. Is there any other way to get entropy from the web?