DNS Problem?

[#1]

I just started getting DNS errors (no response) so had to switch to Google DNS 8.8.8.8

[#2]

I had a DNS problem around 2pm, the internet just suddenly stopped for no reason and then reconnected about 5 mins later, was quite weird.

[#3]

Hi there,
We've checked our monitoring/graphing relating to DNS and can't see any isues.

FYI ReedRichards,
Wondering if it was a disconnection rather than an actual DNS issue? Could be that you reconnected around the time you changed to google DNS?
<img src="http://ccgi.psmith12.plus.com/visradius/generated/image13584435979301.png"/>

[#4]

Adam,
I have expressed concern that both DNS servers appear to be in the same IP subnet and have a common route to them,  Thereby they both share same / similar points of failure.  Is anyone looking at this?  Would there be merit in PN setting up TBB ping graphs for the two PN DNS?

Quote

Tracing route to cdns01.plus.net [212.159.6.9] over a maximum of 30 hops:   1    *        *        *    Request timed out.   2    35 ms    35 ms    35 ms  lo0-central10.pcl-ag01.plus.net [195.166.128.182]   3    35 ms    35 ms    35 ms  link11-central10.pcl-gw01.plus.net [84.93.249.20]   4    35 ms    35 ms    35 ms  xe-0-2-0.pcl-cr01.plus.net [212.159.1.0]   5    35 ms    35 ms    37 ms  po2.pcl-gw01.plus.net [195.166.129.41]   6    36 ms    35 ms    35 ms  vl63.pcl-lb01.plus.net [212.159.2.252]   7    36 ms    36 ms    36 ms  cdns01.plus.net [212.159.6.9] Trace complete. C:\Users\Kevin Johnson\Downloads\Misc\RouterStats>tracert  212.159.6.10 Tracing route to cdns02.plus.net [212.159.6.10] over a maximum of 30 hops:   1    *        *        *    Request timed out.   2    38 ms    40 ms    78 ms  lo0-central10.pcl-ag01.plus.net [195.166.128.182]   3    36 ms    36 ms    35 ms  link3-central10.pcl-gw01.plus.net [84.93.249.4]   4    35 ms    66 ms    34 ms  xe-0-2-0.pcl-cr01.plus.net [212.159.1.0]   5    35 ms    35 ms    35 ms  po2.pcl-gw01.plus.net [195.166.129.41]   6    36 ms    36 ms    36 ms  vl63.pcl-lb01.plus.net [212.159.2.252]   7    36 ms    35 ms    36 ms  cdns02.plus.net [212.159.6.10] Trace complete.

Route to BOTH servers is identical, therefore (on face of the above info)  if any one part of the route fails, access to BOTH servers is inhibited - unless each hop has an alternative route to the target.

Regards,
Kevin

[#5]

Quote from: _Adam_Walker_

Wondering if it was a disconnection rather than an actual DNS issue? Could be that you reconnected around the time you changed to google DNS?

Yes, Adam, I checked the router and the PPP session was dropped and re-established (i.e. no loss of sync).  Sorry, I should have checked before I posted (this type of event is relatively rare for me).

[#6]

@townman
We have DNS servers at more than 1 data centre (there are actually 4 addresses, not 2), these are load-balanced so you'll likely see the less busy ones each time you test.
The 4 addresses are:
212.159.6.9
212.159.6.10
212.159.13.49
212.159.13.50

[#7]

Hi Chris,
Thank you for the additional information.  Though there are more than two DNS servers, it would appear that the DHCP servers are only configured to publish two of them and that pair are in the same subnet.  The only two I have seen (through near 24x7 modem syslog data collection whilst investigating an on-going problem) are 212.159.6.9 & 212.159.6.10.  [Data gathered from 79 establishments of PPP sessions between 8th Jan and 18th Jan on the account linked to ticket 63223897]. 
IMHO the DHCP servers should give a pair of DNS servers which are in distict subnets, each having (as far as pratical) seperate routing paths (for example) 212.159.6.9 & 212.159.13.49 or 212.159.6.10 & 212.159.13.50 - otherwise a failure along the common path renders no access to DNS at all.

[#8]

Quote from: townman

Adam, I have expressed concern that both DNS servers appear to be in the same IP subnet and have a common route to them,  Thereby they both share same / similar points of failure.  Is anyone looking at this?  Would there be merit in PN setting up TBB ping graphs for the two PN DNS? Quote Tracing route to cdns01.plus.net [212.159.6.9] over a maximum of 30 hops:   1     *        *        *     Request timed out.   2    35 ms    35 ms    35 ms  lo0-central10.pcl-ag01.plus.net [195.166.128.182]   3    35 ms    35 ms    35 ms  link11-central10.pcl-gw01.plus.net [84.93.249.20]   4    35 ms    35 ms    35 ms  xe-0-2-0.pcl-cr01.plus.net [212.159.1.0]   5    35 ms    35 ms    37 ms  po2.pcl-gw01.plus.net [195.166.129.41]   6    36 ms    35 ms    35 ms  vl63.pcl-lb01.plus.net [212.159.2.252]   7    36 ms    36 ms    36 ms  cdns01.plus.net [212.159.6.9] Trace complete. C:\Users\Kevin Johnson\Downloads\Misc\RouterStats>tracert  212.159.6.10 Tracing route to cdns02.plus.net [212.159.6.10] over a maximum of 30 hops:   1     *        *        *     Request timed out.   2    38 ms    40 ms    78 ms  lo0-central10.pcl-ag01.plus.net [195.166.128.182]   3    36 ms    36 ms    35 ms  link3-central10.pcl-gw01.plus.net [84.93.249.4]   4    35 ms    66 ms    34 ms  xe-0-2-0.pcl-cr01.plus.net [212.159.1.0]   5    35 ms    35 ms    35 ms  po2.pcl-gw01.plus.net [195.166.129.41]   6    36 ms    36 ms    36 ms  vl63.pcl-lb01.plus.net [212.159.2.252]   7    36 ms    35 ms    36 ms  cdns02.plus.net [212.159.6.10] Trace complete. Route to BOTH servers is identical, therefore (on face of the above info)  if any one part of the route fails, access to BOTH servers is inhibited - unless each hop has an alternative route to the target.

Well, not quite identical, as the third hop is slightly different. And you'll be disappointed to discover that the route to 6.9 is the same as 13.49, and the routes to 6.10 and 13.50 are the same. I remember it's been said elsewhere that they're supposed to be used in the pairs given, 6.9 and 6.10 or 13.49 and 13.50.