I have taken the liberty to tinker with the dns stuff:
I changed the dns data file to:
- make box2 explicit
- point A for 'squeak.org' and 'www.squeak.org' to box4
- point MX for 'squeak.org' to 'box2.squeakfoundation.org'
- point * to box2
As soon as the caches update, everything should be fine again™
These changes look great.
Thanks, Tobias!
Chris
Tobias did some great work yesterday. As the TTL was set at the default we'll know for certain the full effects by UTC 21:40, which will be 24 hours after he posted to the box list about his changes.
I'm sure that when people check the services they will find he did a great job and everything is fine.
In the event that something is left out, I would ask that he go in and make some changes he thinks may be necessary. I'd like to add one suggestion. Please append each line with :3600 to set the TTL to one hour. If something is wrong and it takes another day to see the problem wash out, after that it will be an iteration period of only 3600 seconds. Added to which, Levente has suggested that some services should be left at that TTL, as it assists with analysis when things go wrong. You'd have to ask him further about what that means.
So a line that looked like this:
+klickt.info:85.10.192.171 would look like this:
+klickt.info:85.10.192.171:3600
I'm sure that Tobias got it right the first time and that this is just some brainstorming for the future.
Thank you for your patience, Chris
On 13.03.2014, at 15:56, Chris Cunnington websela@yahoo.com wrote:
Tobias did some great work yesterday. As the TTL was set at the default we'll know for certain the full effects by UTC 21:40, which will be 24 hours after he posted to the box list about his changes.
I'm sure that when people check the services they will find he did a great job and everything is fine.
In the event that something is left out, I would ask that he go in and make some changes he thinks may be necessary. I'd like to add one suggestion. Please append each line with :3600 to set the TTL to one hour. If something is wrong and it takes another day to see the problem wash out, after that it will be an iteration period of only 3600 seconds. Added to which, Levente has suggested that some services should be left at that TTL, as it assists with analysis when things go wrong. You'd have to ask him further about what that means.
So a line that looked like this:
+klickt.info:85.10.192.171 would look like this:
+klickt.info:85.10.192.171:3600
I'm sure that Tobias got it right the first time and that this is just some brainstorming for the future.
We could do that, but reading http://cr.yp.to/djbdns/tinydns-data.html suggests that: a) Tinydns actually picks sensible defaults: “tinydns-data will use default cache times, carefully selected to work well in normal situations.” b) the TTLs are typically different for different entries, as can be seen near the end of the site, where tinydns-format and bind-format is compared. You better don't meddle with those, I think. Best -Tobias
I did a quick circuit around the subdomains and everything looks fine. Unless somebody sees something else that's been dropped, I'd say things are now back to normal. That being the case, I'll turn tomorrow to updating squeak.org with the backlog of requested changes on the webteam list.
I have appreciated your patience and restraint in this matter. Time was. people's reactions could have gone a whole other way. I said as much today talking to Colin.
Chris
box-admins@lists.squeakfoundation.org