@dnandell 

The Deco database system will update if it is rebooted or if a database update is pushed by us.

Please log into the Deco via the web GUI and provide the system logs. We can look at those to track and see what it may be.

Since having the issue have you tried swapping the Deco nodes, and also swapping the Ethernet cable to the modem?

@Tony

FYI, I have brand new CAT6E wiring in my house.  I did not use the cable that came with the X20.  I have a brand new Motorola MB8600 DOCSIS3.1 cable modem.  I purchased the TP-Link Deco X20 to go with the rest of the network upgrade.

Also, it seem the X20 causes the cable modem to lose connection as well at the same time.  It's always during a security database upgrade.  I plugged my old router in for over a day and had no issues.  As soon as I plugged the X20 back in, random disconnects.

Seems like it would be a pretty easy fix to be honest.  If anything, make the security updates manual, if wanted, or on a schedule so they can be done in the night.

Thanks.

@dnandell 

Log into your cable modem. Do you have critical 3 errors in the system logs?

@Tony Can you tell me what this means?

@Tony Can you tell me why you asked your previous questions?

@dnandell 

Here is a good explanation of what the critical 3 errors are, this was taken from another site (pickmymodem.com), not TP-Link:

T3 ( Ranging Request Retries Exhausted )

Explanation: The cable modem has sent 16 Ranging Request (RNG-REQ) messages without receiving a Ranging Response (RNG-RSP) message in reply from the CMTS. The cable modem is therefore resetting its cable interface and restarting the registration process. This typically is caused by noise on the upstream that causes the loss of MAC-layer messages. Noise could also raise the signal-to-noise ratio (SNR) on the upstream to a point where the cable modem’s power level is insufficient to transmit any messages. If the cable modem cannot raise its upstream transmit power level to a level that allows successful communication within the maximum timeout period, it resets its cable interface and restarts the registration process. This error message is DOCSIS event message is R03.0, Ranging Request.

T3 timeouts result when a Cable Modem, which is locked into the downstream, transmits a Ranging Request(RNG-REQ) to the CMTS(Cable Modem Termination System) and 200 milliseconds passes without the CM receiving a Ranging Response(RNG-RSP) from the CMTS. Indeed, maintaining a PING(Packet InterNet Groper) less than 200ms is important to a healthy DOCSIS network. When 10 T3s occur in succession (a period of 20 seconds passes without the CM receiving a Ranging Response, the CM will cease attempting to communicate with the HFC(Hybrid Fiber Optic Coaxial) network and reset its DOCSIS interface. In this case, the RESET statistic on the modem log will increment but not all T3s will cause a reset if communication is restored within 10 T3 periods or 20 seconds. When T3 timeouts occur, but not enough occur in succession to trigger a reset, slow speeds are often a symptom as these timeouts slow down the TCP/IP handshakes necessary to maintain internet connections.

T3 Timeouts are typically caused by Upstream Noise causing Ranging Requests to not be clearly received by the CMTS. T3 problems can often be intermittent in nature as they may be the result of noise originating in any part of the plant that shares the same line card at the CMTS with the subscriber impacted. In fact, T3s can even be caused by noise originating on other neighboring nodes if resources are shared at the headend. Speed issues and intermittent connectivity issues are some of the most common problems encountered on our Trouble Calls.

T3 Timeouts can also be encountered when the plant's noise floor is so severe that the CM cannot overtalk it (SNR is bad and Noise power level is higher than your modem's signal). This can affect a single leg of the plant or even entire nodes. When a high noise floor results in several modems resetting their DOCSIS interface because 10 T3 timeouts have occurred in succession, as we covered above, this is referred to as a noise outage. Noise outages frequently occur node-wide and while some may self-clear, many continue until a Plant Maintenence Technician locates and repairs the offending source of noise. This is why we ALL must be vigilant about preventing noise from entering the plant.

Occurrence: Extremely common in the customer homes

Symptom: Will cause intermittent disconnections. If you see a couple of them, it is fine. If you quite a few of them consecutively, then it will lead to disconnections.

Example:
2017-6-14, 00:05:06 Critical (3) No Ranging Response received - T3 time-out;CM-MAC=00:11:22:c1:db:09;CMTS-MAC=4c:00:82:ee:bd:62;CM-QOS=1.0;CM-VER=3.0

Fixes:
1. Check for loose or Bad cables. This is by far the most neglected and most common fix.
2. Remove extra splitters in house

3. Buy powered amplifier to boots upstream power
4. Buy better coax cables
5. Call for a technician visit to inspect and fix cables outside and inside home (could be rain damage, fire damage, animal damage, construction damage, etc)

Since the X20 has been out, we have not received reports of this happen so when you mentioned the internet constantly dropping, that's why I thought of the crit 3 errors which produce that random connection drop.