DNS

[#1]

I google searched for: isc.org ripe.net dns
and found various things about DNS amplification attacks, so I'm guessing the 20 or 30 second delay is some sort of security feature?
Which finds another domain Plusnet DNS returns SERVFAIL for: dns-oarc.net - The Domain Name System Operations Analysis and Research Center (DNS-OARC).
It can't be a coincidence that isc.org, ripe.net and dns-oarc.net are all large Internet infrastructure related entities.

[#2]

Quote from: ejs

It can't be a coincidence that isc.org, ripe.net and dns-oarc.net are all large Internet infrastructure related entities.

Good point.
But it's still only Plusnet DNS which has the problem.  
Lookup time for ripe.net
root servers == 31 ms
google dns == 46 ms
Plusnet DNS == connection timed out.
Edit 9:30pm
212.159.6.9 has started working for me
212.159.6.10 still has problems plus an additional 4 urls it has problems with.
nominet.org.uk
verisigninc.com
nlnetlabs.nl
kirei.se

[#3]

Some interesting links there ejs & npr thanks. All seemed to have been resolved yesterday evening.

[#4]

This morning, Plusnet's DNS servers are resolving all domains previously identified as problematic, and reverse lookups are working, so the gateway checker works.
But I have noticed that Plusnet's DNS servers now don't respond to type ANY queries (requests for all records), I don't remember if they did before:
[tt]$ dig @212.159.6.9 plus.net ANY
; <<>> DiG 9.9.2-rl.028.23-P2-RedHat-9.9.2-10.P2.fc18 <<>> @212.159.6.9 plus.net ANY
; (1 server found)
;; global options: +cmd
;; connection timed out; no servers could be reached
$ dig @212.159.6.9 plus.net NS
; <<>> DiG 9.9.2-rl.028.23-P2-RedHat-9.9.2-10.P2.fc18 <<>> @212.159.6.9 plus.net NS
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 8786
;; flags: qr rd ra; QUERY: 1, ANSWER: 2, AUTHORITY: 0, ADDITIONAL: 2
;; QUESTION SECTION:
;plus.net. IN NS
;; ANSWER SECTION:
plus.net. 23693 IN NS ns2.force9.net.
plus.net. 23693 IN NS ns1.force9.net.
;; ADDITIONAL SECTION:
ns1.force9.net. 85841 IN A 195.166.128.16
ns2.force9.net. 85841 IN A 195.166.128.17
;; Query time: 19 msec
;; SERVER: 212.159.6.9#53(212.159.6.9)
;; WHEN: Sun May 19 06:49:52 2013
;; MSG SIZE  rcvd: 101
[/tt]

[#5]

Quote from: ejs

But I have noticed that Plusnet's DNS servers now don't respond to type ANY queries (requests for all records), I don't remember if they did before:

Borked. 

[#6]

Going back to the beginning of this thread, when many people saw a DNS latency glitch at 10:00 on 17th May 2013

I question whether SamKnows reporting is showing the correct time.  I have suspected it was adrift since I first got the whitebox, but my connection is so stable that until now I have not been able to correlate any timed events to confirm my suspicion.
Yesterday outside my house an electricity supply pole got damaged, the power company had to come out to replace the pole and the feed cables to the group of houses.  In doing so we had a power cut from almost exactly midday until just before 16:00.
This can be seen on my TBB ping graph -

My SamKnows box doing hourly tests for latency shows this -

The SamKnows reporting has three missing results at 11:00, 12:00, and 13:00, but is back at 14:00.
The TBB ping graph which IS showing the correct time would suggest that the missing three results should be at 13:00, 14:00, and 15:00.
My SamKnows reporting is set to "Timezone : (UTC 0) United Kingdom/London", but the results appear to be 2 hours adrift.
Therefore the apparent DNS glitch on the 17th May was probably at midday rather than 10:00 as indicated !

[#7]

@purleigh my graph shows 10 am also, however! I had bt working on my line from 8am the other day and everything seems in order.

[#8]

@30FTTC06 - What timezone does your SamKnows dashboard say ? (under "Support" => "Your Details")

[#9]

Timezone  (UTC 0) United Kingdom/London

[#10]

To see what is going on, I earlier disconnected from the internet at 11:30 and then reconnected at 12:30.

The corresponding SamKnows graph shows one missing 'on the hour' test, and that is for 10:00 as I suspected.

Therefore my SamKnows reporting graphs are all 2 hours adrift !
Looks like I need to raise yet another issue with SamKnows ! 

[#11]

Good plan, I would drop my connection purleigh just to make sure, but my line is still in a training period. I have noticed things though not quite as they should be, but it all seems to work out in the end. For instance i'll look at a graph from the day before and think to myself no thats not correct! but when i zoom in on the graph a day later everything seems to be as expected, but i also see a lot of joins in the graphs as well where the graph just seems to auto join the gaps together. This makes sense when you take into account some stuff registers on the hour and other stuff every 2 and sometimes 4 hours.
Anyway, all will become clear a bit later on
Sorry i've gone off topic people, but i thought it worth mentioning.

[#12]

Don't know if this is being done because of any of the issues in this thread: http://usertools.plus.net/status/archive/1369050113.htm

[#13]

I think the Sam-thing is reporting times in UTC - Summer time is UTC+1. 
If it takes readings at hourly intervals, is it reporting the beginning or the end of the hour?

[#14]

Quote from: townman

@Matt, Does support know why its was broken? Given that this has been seen before, does support know how to prevent it happening again? Cheers, Kevin

[#15]

This: http://usertools.plus.net/status/archive/1369050113.htm should have fixed it this morning We've upgraded the software and flipped the load balancer which should stop it from happening again.