Got woke up relatively early this morning after a hard night’s toiling to hear from Francois that the application was down. After all of the cursory, standard checks and reboots (Jason had been having a bit of a nightmare here too), the problems still persisted.
And then I tested that we were able to ping from the server – we weren’t. What this has meant in the past, and again here, our upstream providers DNS servers had gone down. This means our server couldn’t talk to the world, and the world couldn’t talk to it.
So, after Jason submited the obligatory (and extremely repetitive) support ticket, I decided to fix this once and for all, and remove our reliance on our upstream providers DNS server and just move to something more dedicated.
In comes OpenDNS – you’ll see two IP addresses at the bottom of that page. I replaced the faulty IP addresses in /etc/resolv.conf to use these (we’re using Ubuntu 8.04 LTS):
nameserver 208.67.220.220
nameserver 208.67.222.222
And after I gave it a ’sudo /etc/init.d/networking restart’ all was well again. You can enter up to three DNS nameservers here, so I’ll probably add our upstream providers one back in at the bottom of the list when it is back up again.
So everything is running again, and I feel more comfortable now, as the DNS servers that we are using are ones that are run by a company who’s core function is to provide these DNS services on these IP addresses – and very little else.
