Read the forum guidelines
Let me start this off by saying that I am livid with Cox support and the inability to get a competent response. I am a Network Engineer with 2 decades of experience and literally spend every day configuring/debugging networks and equipment, from server applications down to the network packet bit level. I have not found any issue with my equipment.
For the past couple weeks my home Cox Internet connection has had random drops and modem disconnects throughout the day. For the past 3 days it has been horrible. My far less technical neighbor confirms "it has gone to $#%7". When it is working, there is no issue and modem signal levels are about as good as you can get. When it drops one of two things happen:
1) Loss of all internet traffic for 15sec to 20min - Modem shows as connected with 42+dBm SNR and little to no errors. Unable to ping the Cox headend gateway from my router. Unable to renew or get a DHCP address from Cox on my router. I can see ARP requests and occasional DHCP responses to other customers, but no response to my DHCP requests. At some point it will suddenly start passing packets again without any changes. Resetting the modem does not fix it.
2) Loss of all internet traffic (modem shows as connected) followed by the modem losing sync a minute or two later and flipping between "provisioning" and "access denied" for the next 2-20min. Sometimes it will provision and appear connected for a few minutes, but I can not get DHCP or internet traffic.
This has to be affecting my entire neighborhood and my neighbor confirms the same dropouts. I finally call in to open a ticket/find out an ETR and Cox Support wants to know what TV I am streaming with... Cox Support wants to know if my computers are hardwired or wireless... Cox Support wants me to reboot my computer... Cox Support wants to know how I know my computer can't get DHCP... None of these have anything to do with my ROUTER connected the the Cox cable modem, not getting DHCP or packets from Cox.
The only answer I have gotten out of Cox is that my modem looks fine and if I agree I can have a Cox Tech come out and "test" my equipment for $75/hr. How do you get a competent response out of Cox Support when the problem is upstream of the modem? This is likely a Cox headend or backhaul issue...
choprboy said: flipping between "provisioning" and "access denied"
That is interesting. Have you recently had a node upgrade? I wonder if the node and/or provisioning on the account is wrong. What do the logs look like on the modem BTW?
I do not know if there has been a node upgrade. The flipping between "provisioning" and "access denied" has been an ongoing thing for the past year or two when there are outages, not specific to the latest drops. It would run perfectly fine for a few weeks to a couple months, then there would be one or more large bursts of correctable/uncorrectable errors (i.e. go for a month with only a few dozen to a few hundred errors, then have 20,000 errors) and the connection would be down for a bit before recovering. I chalked those up to splicing/maintenance work being done.
The modem is an Asus CM-16 16x4. Like basically all cable modems, it is neutered and has little to no actual debugging utility. It shows registration/connected state and up/down channels/signal level. There are no logs or advanced interface that I have been able to find. I have had an "Internet Essentials 30/3" plan, not doing a lot of bandwidth just basic internet and VPN for work, though looking at my account yesterday it was apparently that was changed to a 50/3 plan by Cox at some point.
Something has changed overnight - both to my Cox account and a new connection symptom this morning. I started the today with 6-8% continuous packet loss to the Cox headend/gateway and points beyond. Previously I had no packet loss up until the point of all traffic stopping. As of yesterday (I don't recall before then) I had 16 bonded downstream channels (all QAM256, 42-43dB SNR) and 1 bonded upstream channel (ATDMA 5MK/s 45dB). I rebooted the modem yet again and it came back up with 4 bonded upstream channels now and 1-2% packet loss. So it appears (at least this new symptom) to be related to 1 of the 4 upstream channels, probably the 23.5MHz that was the single previous channel.