Has Plusnet server 212.159.6.9 just gone down ?

[#1]

My router (for at least a decade) has used the server at 212.159.6.9 as a ping target, to act as a connection watchdog to detect when my broadband connection is down.

I've just discovered that 212.159.6.9 is currently not responding to pings, but it's sister 212.159.6.10 is working OK

Is it just me seeing this ?,  and if not then perhaps someone at Plusnet might want to check that the server is OK ?

@bobpullen 

@willcutforth 

[#2]

---

@Anonymous wrote:

Is it just me seeing this ?

---

I'm seeing the same.

[#3]

Likewise, but its still serving DNS requests.

@Anonymous has it just started to not respond or has it possibly been like that for some time ?

[#4]

I've never seen it not working before.

It was definitely OK yesterday, as I spent hours fiddling with my upgraded router, and would have noticed the big red banner indicating a failure of gateway monitoring.

Likewise this morning I was looking at it, while talking to Plusnet tech support, and it was pinging OK then.

I believe this loss of ping only started sometime this afternoon

[#5]

@Anonymous
Earlier this afternoon I happened to ping the DNS servers auto-configured on my connection, and the four recommended here:
https://community.plus.net/t5/Community-Announcements/Important-information-about-our-DNS-server-addresses/td-p/703337

Results:
212.159.13.51: No response
212.159.13.52: No response

212.159.6.9: responded
212.159.6.10: responded
212.159.13.49: No response
212.159.13.50: No response

After reading your message, 212.159.6.9 no longer responds.

Regards,
Alistair

[#6]

@Anonymous Thanks for confirming that. I know there has been work on the DNS servers but that was completed 3 or 4 weeks ago so probably not connected. Bob will know about it though

[#7]

Seems it went down;

Wed Jul 12 14:17:05 2023		2:27:45
Tue May 09 21:44:42 2023		63 days 16:32:23
Sun May 07 16:16:35 2023		2 days 05:28:07

[#8]

I don't know if it connected but after the last Hub Two Firmware upgrade to v0.10.00.04201-PN the DNS servers have changed from 212.159.6.9 and 212.159.6.10 to Primary DNS: 212.159.13.52  and Secondary DNS: 212.159.13.51, neither of these hosts respond to ping, 212.159.6.10 still does for me.

I set my pfSense to pingbox1.thinkbroadband.com [80.249.99.164], makes more sense to ping something off network.

[#9]

---

@Dan_the_Van wrote:

I set my pfSense to pingbox1.thinkbroadband.com [80.249.99.164], makes more sense to ping something off network

---

Yes I can see your point,

I was working on the principle of using the absolute nearest ping endpoint so I can use the gateway monitor graph as an indication of baseline latency for my connection.

I think I used to use "ntp.plus.net" as I figured that a local NTP time server should have low latency and low jitter,  but now that URL resolves to "ntp.r53-gslb.vision.bt.com (213.123.16.85)"  which I assume is off Plusnet's local network.

[#10]

---

@MisterW wrote:

@Anonymous Thanks for confirming that. I know there has been work on the DNS servers but that was completed 3 or 4 weeks ago so probably not connected. Bob will know about it though

---

We are mid-migration to a new DNS platform that in the short term will see RADIUS assign two new virtual IP's as part of the PPP negotiation (212.159.13.51/52). These IP's currently point to the new platform.

@Dan_the_Van wrote:

I don't know if it connected but after the last Hub Two Firmware upgrade to v0.10.00.04201-PN the DNS servers have changed from 212.159.6.9 and 212.159.6.10 to Primary DNS: 212.159.13.52  and Secondary DNS: 212.159.13.51, neither of these hosts respond to ping, 212.159.6.10 still does for me.

The firmware itself has nothing to do with it, however because the upgrade invokes a hub reboot, it has accelerated the migration of customers over to the IPs mentioned above.

I haven't been that close to the work, however I understand that the next steps were to point the longstanding virtual IPs (212.159.6.9/10 and 212.159.13.49/50) to the new platform.

@G8DHE wrote:

Seems it went down;

Wed Jul 12 14:17:05 2023

2:27:45

My guess is that 212.159.6.9 was pointed to the new platform around this time, and the new platform is not comnfigured to respond to ICMP traffic (unsure if intentional or an oversight, probably the former).

@dave is the man to confirm though.

[#11]

Yeah as Bob says we're moving on to a brand new DNS platform with a set of brand new servers. To aid the migration across we starting assigning 212.159.13.51/52 when people connected last month, if there were any problems then they would be easy to spot by doing this, we were able to go slowly and it was easy to revert back.

We also split DNS and NTP up as they were on the same IP addresses and servers, they are now separate.

The new servers though don't respond to ping which is why you're seeing this, this is intentional. We've started to repoint the existing IP addresses too; 212.159.13/49/50 went a week or so ago, 212.159.6.9 went yesterday afternoon and 212.159.6.10 is planned to move on Tuesday afternoon. We needed to repoint the IPs as there will be people that have hardcoded the IPs in their routers and devices at some point and we didn't want to cause them any issues. The fortunate timing of the Hub2 upgrade caused anyone with a Hub2 to be moved over to the new DNS platform as the upgrade forced a disconnection so picked up the new IPs when it came back up.

Most people are moved over now and once we've moved 212.159.6.10 next week anyone that was using the old DNS servers will be on the new servers, we'll also then revert the IPs back to what they were the next time people disconnect and reconnect which will allow us to keep the 212.159.13.51/52 IPs for the next time we want to do major work on DNS.

[#12]

@dave 

---

@Anonymous wrote:

My router (for at least a decade) has used the server at 212.159.6.9 as a ping target, to act as a connection watchdog to detect when my broadband connection is down.

I've just discovered that 212.159.6.9 is currently not responding to pings, but it's sister 212.159.6.10 is working OK

---

---

@dave wrote:

The new servers though don't respond to ping which is why you're seeing this, this is intentional. We've started to repoint the existing IP addresses too; 212.159.13/49/50 went a week or so ago, 212.159.6.9 went yesterday afternoon and 212.159.6.10 is planned to move on Tuesday afternoon.

---

---

@Anonymous wrote:

I was ... ... using the absolute nearest ping endpoint so I can use the gateway monitor graph as an indication of baseline latency for my connection.

I think I used to use "ntp.plus.net" as I figured that a local NTP time server should have low latency and low jitter,  but now that URL resolves to "ntp.r53-gslb.vision.bt.com (213.123.16.85)"  which I assume is off Plusnet's local network.

---

---

@dave wrote: We needed to repoint the IPs as there will be people that have hardcoded the IPs in their routers and devices at some point and we didn't want to cause them any issues.

---

Yesterday when  212.159.6.9  stopped responding to pings and my router's WAN monitoring watchdog constantly timed out (causing the router to repeatedly try and reboot the PPPoE connection in an attempt re-establish a connection that did respond to pings),  I swapped my settings to 212.159.6.10  to keep my connection running.

So you're now saying that  212.159.6.10  will also stop responding to pings on Tuesday afternoon.

So my question is - what NEW server URL and IP address would you recommend that I hardcode into my router, that will respond to pings,  that will have the lowest latency and jitter  assuming that my connection endpoint is Telehouse in London.

[#13]

ntp.plus.net is now within the BT network and uses the same servers as BT customers (it's what we used as source for NTP anyway so removes a point of failure). It also doesn't respond to ping.

212.159.6.10 will stop responding on Tuesday afternoon.

In terms of suggesting a new IP I think to be honest I'd say Google's 8.8.8.8 might be the best. There aren't many IPs within the Plusnet broadband infrastructure that will respond to ping and of the ones we do have some of them might get decommissioned within the next 12 months so you'd need to change again.

[#14]

Ok thanks @dave 

I'll have to consider which non-Plusnet ping target I will use going forwards.

I currently have a Hurricane Electric 6in4 tunnel on my router to provide IPv6 network connectivity, their servers appears to be located in the same building as yours, so I'll go and experiment with pinging off their IPv4 tunnel endpoint.

[#15]

Having done some ping testing using the 'mtr' command (Linux equivalent of 'tracert')

• Plusnet 212.159.6.10 had the lowest latency and lowest jitter
• The Hurricane Electric tunnel endpoint also had low latency but frequent high jitter (+40ms)
• Google DNS had the next lowest latency but regular +15ms jitters
• Cloudflare DNS was typically 0.4ms slower than Google DNS, but very little jitter.
• "pingbox1.thinkbroadband.com" (suggested by @Dan_the_Van) had similar latency as Cloudflare, but jitter up to +26ms

So for now, I've set my router's ping target to Cloudflare 1.1.1.1,  and I'll see how it goes.