Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Fezster]

Yes, already disabled gateway monitoring action. And also disabled monitoring bogon networks (which exacerbates the disconnect).

However, switching back to pfsense, didn't take long to see the packet loss / ping spike.

[Fezster]

This is driving me mad.

I said yesterday I saw a spike within an hour of using pfsense (with SMP disabled). So I switched back to OPNsense - which I'd been happily running for 3 days with no spikes and no packet loss - and guess what? I saw a spike/packet loss/gateway disconnect within a couple of hours (!).

Coincidence? Perhaps. Maybe some kind of strange cutover issue between the 2 virtual machines (I run them both on an esxi server).

So Im repeating my pfsense test today. This time, I rebooted my esxi box, so it should be a clean run. Will update later.

[PhilipD]

Hi

I've been using pfSense 2.4.5 on metal with 4 cores enabled and not needed any workarounds, so I guess lucky in that respect, but if affected then p1 release is due soon and fixes it.  I've no packet loss or latency issues on the BQM graph or gateway drops, but then I'm also not with Virgin media, which on the charts I've seen for them, always look poor and jagged.

As your troubling shooting shows sometimes these things are hard to pin down.

If it is only spikes on pfSense and the dropping of the gateway is resolved are those spikes causing a problem?  Remember that the monitoring is only testing pings, which could just be getting a lower priority on pfSense internally in its code than maybe OPNSense.  You could try turning on QoS and putting pings into real-time traffic queue to see if that results in a better chart.  Also if pfSense is only running on 1 core for the work around, that might be enough to affect the ping responses as well.

Regards

Phil

[Fezster]

Hi - thanks for the response.

I think this does seem to be very much a Virgin Media specific issue, from what I've been reading.

I do have QoS already turned on to prioritise pings - but you are right, single core performance could be causing an issue.

However, it is quite obvious and specific when this specific issue occurs. You suddenly get a very high ping spike + packet loss and it always lasts a specific amount of time (around 10 seconds).

[Alex Atkin UK]

I didn't think I had any issues, but considering the terrible experience I had with trialing Stadia yesterday I'm starting to wonder.

I have seen the odd latency spike, but it doesn't seem frequent enough to be this issue, or is it?

[kerman19]

Hello!

I just wanted to chime in here and say that I too (since around early Feb) have had issue with pfsense/opnsense and Virgin Media. It was at the point where I've pretty much turned every single damn feature off to try and resolve it, naturally none of this worked.

My symptoms are(were) pretty much exactly as described in this thread.

An example of my shoddy connection can be found here (on a good day).
https://www.thinkbroadband.com/broadband/monitoring/quality/share/3c940bc31113361aa9a14f8d553109a043597235-21-05-2020

And this on a generally bad day, when I'm using it the most.
https://www.thinkbroadband.com/broadband/monitoring/quality/share/127ebfdc885ffbd30e9e7796d467fe34f4412d8d-14-05-2020

Now this is today
https://www.thinkbroadband.com/broadband/monitoring/quality/share/26214d30be6ee15756204e05efee1113f9bbfdcd-29-05-2020

And yesterday
https://www.thinkbroadband.com/broadband/monitoring/quality/share/aaf09c1a9076591555eda9f3d0fb8537910521d6-28-05-2020

So the fix, at least for me was that by default unbound operates in resolver mode, great, that's what we want! Turns out that it completely borks my virgin connection every time I want to use it, and I think this seems to pair with a firmware update that we took around the time this started happening.

So, for the last two days I've been using forwarding mode (with prefetch off) to 1.1.1.1 and OpenDNS and it's been faultless, the only spikes in the graphs are where I've been downloading at my full 350mbps for extended periods for which I expect latency to increase.

Hope this helps some of you guys with Virgin Connections.

Edit: Fixed the images, in my haste I didn't notice they were all the same!

[underzone]

Check your post mate. All the images are the same:
https://www.thinkbroadband.com/broadband/monitoring/quality/share/thumb/7afe1defa532e5c5ee7b6984c474df360d2340ee.png

[Fezster]

Yes - your pics are all the same. Pls check.

Thanks for the tip - let me try and do the same with unbound.

[Ronski]

I would like to try this as well, not completely clear on what I need to do though.

Is it just a case of disabling DNS Resolver and enabling DNS forwarder?

Anything else that needs doing?

[kerman19]

Yeah, just put it in forwarding mode with some decent (non virgin) DNS servers. I also turned off prefetch (opnsense terminology).

I’ve attached a couple of images from opnsense with the appropriate settings on them.

[Alex Atkin UK]

Yeah, just put it in forwarding mode with some decent (non virgin) DNS servers. I also turned off prefetch (opnsense terminology).

I’ve attached a couple of images from opnsense with the appropriate settings on them.

Actually it depends, you can use dnsmasq or unbound for forwarding.  I can imagine dnsmasq being less problematic as unbound does have its quirks sometimes, it seems to take longer to restart when the firewall reloads for example.

Any particular reason you turned off prefetch?  It should make things faster I believe?

[kerman19]

Actually it depends, you can use dnsmasq or unbound for forwarding.  I can imagine dnsmasq being less problematic as unbound does have its quirks sometimes, it seems to take longer to restart when the firewall reloads for example.

Any particular reason you turned off prefetch?  It should make things faster I believe?

I’d received a modest improvement in previous testing with my connection quality by turning prefetch off, so it is less about improving DNS performance and more about preventing my virgin connection from constantly falling over.

In an ideal world I’d have it in resolver mode with prefetch enabled but in that world virgin media wouldn’t be my ISP, unfortunately I've got more chance of winning the lottery than getting FTTP in my area.

[Fezster]

@Kerman19 - Thank you for the tip. I think it may well have resolved the issue. I'll know for sure during the week when I use my connection constantly for Zoom calls.

Easy to switch in pfsense.

1. Go to Services -> DNS Resolver
2. Tick "Enable Forwarding Mode"
3. Optionally enable "Use SSL/TLS for outgoing DNS Queries to Forwarding Servers"

[Ronski]

Thanks, I've now done that, we#ll see what happens.

There is also a DNS Forwarder service, I thought it was that which was being referred to.

[kerman19]

@Kerman19 - Thank you for the tip. I think it may well have resolved the issue. I'll know for sure during the week when I use my connection constantly for Zoom calls.

Easy to switch in pfsense.

1. Go to Services -> DNS Resolver
2. Tick "Enable Forwarding Mode"
3. Optionally enable "Use SSL/TLS for outgoing DNS Queries to Forwarding Servers"

I have my fingers crossed for you!