Also for those interested:
http://www.intodns.com/trollkingdom.net
That's the DNS report now. Just so you can see that DNS issues can be a bitch to fix and are rather complicated. Everything is working now though.
The three "warning signs" don't apply to us and are nothing to worry about.
Single point of failure in regard to distributed subnets is fine -- it just implies that we don't have our nameservers set up across different servers. Since both of our nameservers are in the same physical location, having them distributed across different subnets has no real-world benefit.
SOA EXPIRE tells servers how long to wait before dumping their cached DNS info, in the case that they can't refresh. This has nothing to do with making DNS changes, and doesn't affect how long changes take to propagate. This assumes a catastrophic fault at your NS. I think 3600000 is longer than RFC recommendations but I don't think it really matters and I can't be bothered to change it.
Before all of this crap went down we had failures in the A Records and the GLUE (the latter being down to doubled up IP's on NS1 and NS2) but it's all working and so the erratic downtime for different pockets of the membership should be done and dusted. Well, I live in hope anyway.
http://www.intodns.com/trollkingdom.net
That's the DNS report now. Just so you can see that DNS issues can be a bitch to fix and are rather complicated. Everything is working now though.
The three "warning signs" don't apply to us and are nothing to worry about.
Single point of failure in regard to distributed subnets is fine -- it just implies that we don't have our nameservers set up across different servers. Since both of our nameservers are in the same physical location, having them distributed across different subnets has no real-world benefit.
SOA EXPIRE tells servers how long to wait before dumping their cached DNS info, in the case that they can't refresh. This has nothing to do with making DNS changes, and doesn't affect how long changes take to propagate. This assumes a catastrophic fault at your NS. I think 3600000 is longer than RFC recommendations but I don't think it really matters and I can't be bothered to change it.
Before all of this crap went down we had failures in the A Records and the GLUE (the latter being down to doubled up IP's on NS1 and NS2) but it's all working and so the erratic downtime for different pockets of the membership should be done and dusted. Well, I live in hope anyway.
