Dropping packets, how to get it resolved?

[brman]

I have ongoing issues with dropped packets. I am pretty sure it is virgin network, not my local connection although to keep VM happy I have rebooted my hub and done a network test from the website, no problems shown but the problem persists. I also called the helpdesk and was (predictably) told they could not see anything wrong and could not do anything.

So, not a situation I can tolerate so how to escalate?

The issue in more detail: On and off for the last month or so I have see a lot of dropped packets, normally late afternoon. Not normally enough to completely kill my internet (although sometimes it does) but enough to kill any voip call or vnc connection.  The issue generally lasts for about an hour and the clears up for the rest of the evening.

I run a quality monitor from thinkbroadband so here are some example days:

Today:

https://www.thinkbroadband.com/broadband/monitoring/quality/share/2eecff228b55e0a695642c82cf41634284... 

Tuesday:

https://www.thinkbroadband.com/broadband/monitoring/quality/share/e96317b6f1c5455c69ce1c651beab7b00d... 

It would not be such an issue except (like a lot of people) I am working from home so need a reliable internet. This is not that.... 😞

Oh, and in case someone asks, no it is not an issue with my local network. I run a firebrick router (VM hub in modem mode) with failover to a 4G modem. That all works 100% as expected.

[Chris_W1]

HI brman, thanks for the message, we have looked into this and cannot see anything wrong, the issue may lie with the other router and we cannot advise on 3rd party equipment, if you are able to change to router mode we may be able to do some tests. Chris. 

[legacy1]

If you go to 192.168.100.1 to view the hubs stats for the upstream bit does it show any T# timeouts?

Can you ping -t 194.168.4.100 when you get packet loss to see if you get packet loss there.

[brman]

Hi Chris,
Sorry for the slow reply. For some reason I only just got the email notification of these replies.

I'll try to find time over the weekend to do as you suggest and report back.

[brman]

legacy1: A good question. I will have to wait until it fails again to test but I will try that.

I have just noticed something a bit odd though. pings to anything (wired) on my lan gives <1ms as expected. I have tried both pc->switch->device and pc->switch->firebrick router->device with the same results. So (at the moment at least) nothing odd about my lan switch and router.

However if I ping 192.168.100.1 then it consistantly averages around 15ms. I can't remember trying this before but an extra 15ms just to get to the VM modem sounds a bit odd to me. Especially as a ping to google.com is only about 19ms average.
I do have some T3 timeouts (30 to 40ish on all 4 channels) but no T1,2 or 4. Looking in the network log all but one of the recent "No Ranging Response received - T3 time-out" messages coincide with times my ping monitor showed significant (ie 10 to 30%) packet loss.

[Robert_P]

Thanks for getting back to us brman, we've had a look at your account and connection and everything does seem to be in spec and range. This is based on the time since the last reboot which was very recent from the time of this post.

If you could monitor it over the next 24 hours and we can have another look to see if anything changes.

Thanks

Rob

[brman]

As of this morning the hub is back in router mode so I will monitor it and report back on how it goes. 

[brman]

I had a secret hope that going to router mode on the hub would fix this problem as it forces a wan ip address change (and hence maybe some routing changes?). Unfortunately not the case.....

https://www.thinkbroadband.com/broadband/monitoring/quality/share/6604e3b8b0ea9688ccb0b39825ffd0745f...

As you will see, the monitor starts from when I changed to router mode this morning (ie new ip address) but lots of dropped packets from 3.45pm ish onwards for about an hour. For much of that time no-one was actually in the house and only one device (windows pc) was connected but in standby. Not that it matters but I didn't even have my lan router (firebrick) or any switches connected.

So, I think I am right in saying that that proves the issue is with either the vm hub or vm network.

Looking at the hub statistics, I have some pre-RS errors (max 32) but no post-RS errors. On the upstream I have 7 to 13 T3 timeouts but no T1, T2, T3. The log showing a couple of "No Ranging Response received - T3 time-out" as before. I am not sure what that tells me.

[brman]

and it is off again.....

https://www.thinkbroadband.com/broadband/monitoring/quality/share/58149ae10b60bb5ac9d6155d87c2fd35da... 

Unfortunately I was on a work call so could not check pings but I have now set up a ping graph from my firebrick to 194.168.4.100 so next time I can see if they are getting that far.

What I do know is that pings both ways (ie thinkbroadband to VM router) and my LAN firebrick to google.com are both dropping. Here is my firebrick to google.com for this afternoon.....

https://www.dropbox.com/s/uwa1ejaneje681a/brick.png?dl=0 

as you can see, not a lot of traffic at the time (a voip call and VNC link, until they were both killed by the dropped packets).

[brman]

To add a bit more evidence.... there was a very brief dropout after the previous graph. This time when I had my firebrick pinging the dns server and the hub.

https://www.dropbox.com/s/c332ezrk5jtuhxe/brick2.png?dl=0 

vmdnsserver is 194.168.4.100, ie VMs primary dns server.

VMHub is 192.168.0.1, ie my VM hub.

As you will see, no problem with pings from the firebrick to the hub so I guess this shows the hub has not locked up? But packets to the dns server are dropping.  Not done at the same time as the dropouts but in case it helps, a tracert from my PC....

https://www.dropbox.com/s/ykv770obh0momli/tracert.png?dl=0