0

I'm running Nginx on an Amazon EC2 micro instance, with some some backend node processes and mongodb.

Recently, I had about 5k hits in about 40 hours, and after a little while,the SSL site hung on request, with chrome returning a timeout.

I restarted the VM, and that did let the site load, but now it still hangs upon requesting the SSL site. Oddly enough, Chrome also says that the site's certificate is invalid, but far as I am aware the certificate is (and was before this all happened) valid.

Any ideas on why SSL is hanging?

(I did not edit configuration before or after the restart).

2 Answers 2

1

I would be very wary of doing anything CPU intensive on a micro instance. Since CPU resources are on-demand, if/when you you do something demanding for a substantial amount of time (doesn't have to be that long), your processing power will be throttled. You will be able to see this as stolen CPU cycles when running vmstat or top (%st is the column to look for).

So next time when nginx is being unresponsive, check to see if you are getting stolen cycles.

On an EC2 small instance (or better), the CPU resources are guaranteed.

1
  • It shouldn't be too processor intensive. The CPU/Memory usage has been pretty stable. The load average is at 0.00 or 0.2 max in top. It's only a problem with the Nginx SSL. Everything else is responsive.
    – CodeJoust
    Jan 9, 2012 at 19:27
0

Does /dev/urandom exist? because if it doesn't, SSL Handshake/Negotiation block reads on /dev/random instead.. which is bad for a micro instance really. We avoided this issue before by using Ubuntu 11.10 on EC2 micro instance.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .