Hello, I'm using Rsyslog on Ubuntu 12.04 on Amazon's EC2. We're using rsyslog 4.2.0.
Our web servers log to local rsyslog processes, which then send the logs on to a central logging server. Recently, the central logging server was rebooted. After it came back up, no more logging appeared from our application nodes, until I rebooted rsyslog on the application nodes. I'd like logging to automatically start appearing on the central server after reboots in the future. Here's what I suspect happened: - We'd specified the central logging server so that it resolves to one of amazon's special dual-DNS-horizon host names that are associated with Elastic IPs. Hosts outside of Amazon see a static IP, while hosts inside Amazon (like our app nodes) resolve the host name to a IP address which changes when the server reboots. It appears that rsyslog was aggressive about caching this DNS resolution, and didn't check again to see that the host name of the target server was resolving to a new IP. I could update the application nodes to use the "Elastic IP" of the central logging server. This has the advantage that it never changes, but I think the routing may be less direct than using the dynamic internal IPs. What advice do you to address this situation? Is there a way to get this version of Rsyslog to respect the TTL of the target rsyslog host? Thanks! Mark _______________________________________________ rsyslog mailing list http://lists.adiscon.net/mailman/listinfo/rsyslog http://www.rsyslog.com/professional-services/ What's up with rsyslog? Follow https://twitter.com/rgerhards
