Menu
Reply
  • 11.38K
  • 254
  • 2.12K
Superuser
Superuser
1,230 Views
Message 3941 of 4,479
Flag for a moderator

Re: Hub 3 / Compal CH7465-LG (TG2492LG) and CGNV4 Latency Cause


@Sephiroth wrote:

For what my opinion is worth, a subjective view in this case is not invalid.

 


Thanks - Too many of us it's worth a lot and so is any opinion irrespective of technical aptitude but still depends whether the OP is willing to "objectively" pursue with further diagnostics/effort?  Clearly I failed to give a balanced post so edited to attempt some "balance"....  Smiley Wink

Regards Tony
"Life is a Binary Inspired Turing Computed Hologram"(don't PM or @Mention me - in case ignoring you offends)
0 Kudos
  • 27
  • 3
  • 0
dusktilldawn6
On our wavelength
926 Views
Message 3942 of 4,479
Flag for a moderator

Re: Hub 3 / Compal CH7465-LG (TG2492LG) and CGNV4 Latency Cause

Quick question I'm on the old firmware still 9111.6v im a new customer since 3rd may 2018 ,im on the 350 virgin package not an upgrade ,how do i get the new firmware thanks

0 Kudos
  • 41
  • 1
  • 7
imzrex
Dialled in
905 Views
Message 3943 of 4,479
Flag for a moderator

Re: Hub 3 / Compal CH7465-LG (TG2492LG) and CGNV4 Latency Cause

You need to factory reset the hub for them to push the new firmware to the hub
0 Kudos
  • 27
  • 3
  • 0
dusktilldawn6
On our wavelength
894 Views
Message 3944 of 4,479
Flag for a moderator

Re: Hub 3 / Compal CH7465-LG (TG2492LG) and CGNV4 Latency Cause

Done that on previous hub and faulty hub showed software failed download or install under log,and engineer replaced faulty hub ,and have hard reset current hub 3,i signed up for 350 new connection not upgrade and  i see upgraded people get it instantly ,do i just wait, or request it or any thing i can do ,any input welcome thanks

0 Kudos
  • 6
  • 0
  • 0
mercuryin
Joining in
891 Views
Message 3945 of 4,479
Flag for a moderator

Re: Hub 3 / Compal CH7465-LG (TG2492LG) and CGNV4 Latency Cause

For me these spikes are still a big problem. The problem is just fixed partially.

 

capture.jpeg

0 Kudos
  • 149
  • 0
  • 15
nad223
Dialled in
797 Views
Message 3946 of 4,479
Flag for a moderator

Re: Hub 3 / Compal CH7465-LG (TG2492LG) and CGNV4 Latency Cause

907ada1fccd191f40bd06b883062c10106e8c1ba

Software version: 9.1.116.603
0 Kudos
Highlighted
  • 102
  • 2
  • 91
Datalink
Up to speed
658 Views
Message 3947 of 4,479
Flag for a moderator

Re: Hub 3 / Compal CH7465-LG (TG2492LG) and CGNV4 Latency Cause

@mercuryin, looks like you have packet loss from the second hop all the way to your selected target. Assuming that you have the modem running in router mode, or, the modem running in modem mode with a follow on router, the second hop is the CMTS. The 10.166.192.1 address is unusual as it falls within the private addres range, but, if that is what Virgin Media is using for the CMTS address, ok, it is what it is. Use that CMTS address as the ping target, not bbc.co.uk. For the purposes of detecting ping spikes from a Puma 6 modem, you don't have to go any further than the CMTS. In fact, if you do, that will usually mask the issue and result in a final plot that looks better than it really is.

So, for the present time, use 10.166.192.1. Drop the ping interval by overwriting the default time in the interval window down to 0.1 or 0.05 seconds or possibly further if you prefer. Usually the response time to the CMTS varies between 0.008 to 0.013 seconds, give or take a few milli-seconds. Drop the ping plot time down to 60 seconds. Right click on the plot and select 60 seconds. That should easily show if you still have issues with ICMP ping spikes that are attributable to the modem. Look for packet loss as well. Its possible for Pingplotter to generate false packet loss indications for the CMTS hop. To confirm whether or not you are actually seeing packet loss, run a command line ping and let that run for a few hours: ping -t 10.166.192.1

Use Control^C to bail out of the ping test and look at the results for any packet loss. If there is no packet loss in the ping test, then for some reason, in your particular case Pingplotter is generating false packet loss indications. Disregard any packet loss indications from hop two if that is the case.

If you don't see any ping spikes as the plot is rolling along, then the ping spikes that you see in your current plot are not attributable to the modem. If your modem is running the newer firmware version that would make sense.

Note that Pingplotter uses data averaging but doesn't advertise that fact. With low ping intervals you end up with a large amount of ping data. As you go up in plot time scales, from 60 seconds to 5 min, 10 min, 24 hours, etc, you are compressing more data into a fixed horizontal size display. Pingplotter has chosen to average the data for the number of data points which lie within a horizontal pixel or pixel range instead of preserving the low and high data points. As you go up in time plots that data averaging makes the plot look pretty good, even if you have high latency points in the data, so, that forces you to use the 60 second and perhaps 5 minute display to see the high latency points on the plot.

Pingplotter does display more than one plot. If you click or double click on the hop #2 IP address, that will bring up a second plot, stacked vertically on the display. That will allow you to compare the 2nd and final hop for latency spikes. If they are generated by the modem, you will usually see a one to one latency spike correspondence between the two plots. If you don't have that, then the latency spikes at the target are not caused by the modem.

One setting that is worth changing is to display the MAX return time in the text data. Your max plot times and max text data don't match. I assume that's due to the fact that the MAX time column is not displayed. Right click on the title bar for the columns, select MAX to display the maximum times and drag that column to the right so that it sits beside the MIN column. With that column displayed, you can see the maximum response times across all hops and easily determine if ping spikes from the modem is still an issue.

If you happen to be running a BQM as well, stop the BQM and set the modem so that it does not respond to WAN pings while you're running a Pingplotter test.

I'd be interested in seeing another post, using the CMTS as the target instead of bbc.co.uk, just to see what the end result looks like.

Hope this helps.

0 Kudos
  • 2.06K
  • 130
  • 503
cje85
Problem sorter
645 Views
Message 3948 of 4,479
Flag for a moderator

Re: Hub 3 / Compal CH7465-LG (TG2492LG) and CGNV4 Latency Cause

It's not possible to ping the addresses starting with 10 that VM uses for the CMTS.

 

My first hop (after router) is:

10.250.64.1 (10.250.64.1) 14.126 ms 14.561 ms 14.851 ms

 

But pinging it does not work:

PING 10.250.64.1 (10.250.64.1): 56 data bytes

--- 10.250.64.1 ping statistics ---
5 packets transmitted, 0 packets received, 100% packet loss

0 Kudos
  • 102
  • 2
  • 91
Datalink
Up to speed
635 Views
Message 3949 of 4,479
Flag for a moderator

Re: Hub 3 / Compal CH7465-LG (TG2492LG) and CGNV4 Latency Cause

Very interesting.  So it shows up in the pingplotter output, but you can't ping the CMTS with a command line ping?  In that case, ping the next hop IP address.  That's not entirely ideal, but it does work for the purposes of detecting high latency returns thru the modem. 

0 Kudos
  • 21.74K
  • 619
  • 3.61K
Sephiroth
Alessandro Volta
591 Views
Message 3950 of 4,479
Flag for a moderator

Re: Hub 3 / Compal CH7465-LG (TG2492LG) and CGNV4 Latency Cause

Just to explain about the 10.xx address.

When the modem last initialised, it was a DHCP client to the CMTS and was allocated the 10.xx IP address as a client of the CMTS. Just to square the circle, the modem, acting then as a DHCP server to the router, binds with the router on 192.168.100.20 (usually) and sets up the gateway and WAN IP address.

There is no path from a router's client to the 10.xx address which is only visible to the modem.


Seph - ( DEFROCKED - My advice is at your risk)

0 Kudos