All five kp115 smart plugs went offline according to Sense but are still really on

I was waiting for someone to post becauase I couldn’t post 3 consecutive in a row. As an update to my tech support ticket with TP-Link, it’s currently being looked at by 3 engineers and they appear to be monitoring this thread and have asked me for pcaps and additional information. They seem interested in fixing the issue.

Welcome to the club Kevin!

5 Likes

This is making me a bit paranoid. With the exception of a single test plug, I have all my plugs configured for local control only. I just added a firewall rule to back that up with. Two immediate things to note:

  • The plugs continue to call home even though they are set for local only. This, in a word, is bad.
  • The plugs do not honor the NTP setting from DHCP. Also bad.

No such things as coincidences :man_shrugging:t3:

Installed a HS300 at night. Plugged in my HP 8740 Printer. Next morning my HP8740 was busted. Scanner error.

I wondered immediately. But then said it had to be coincidence.

Later that day saw the strange spike on the HP Printer.

Started wondering more, posted spike info/pic on forum without mentioning the broken Officejet. Several responses about spikes but nothing else.

After reading of other bricked devices here, now wondering more.

And in infinite HP wisdom, the Officejet will not print while scanner (bed) is in error state. So I’m hosed after just spending and installing $175 worth of new ink cartridges, not to mention cost of new Officejet. Why do I bet sky high prices and low availability with microchip shortage?

l changed the DHCP lease to 30 days, also changed the Password on my Kasa cloud account but right on schedule, all 5 KP115’s went N/A EXACTLY 5 days later.

I’m still awaiting a response from TP-Link. I have submitted all the data they asked for. I will update when they reply.

2 Likes

I’m starting a long term packet capture on my test plug to see if I can catch it doing anything unexpected on a 5 day interval.

My current hypothesis (just looking at it logically) seems that there is a bug in the firmware of these plugs that is negatively reacting (crashing) due to a recurring unrelated network message/event that occurs on my network.

We now can probably rule out DHCP now. But what else might fire off a message or series of packets that the KP115’s just can’t handle and they lock up from further local network Sense messages?

There is a lot of stuff on my network: Two MacOS servers, Router, three WIFI APs, Network Switches, VOIP adaptors, several laptops, iPhones, iPads, 50+ IoT devices (light bulbs, switches, locks, etc), Media Players, Smart TVs, Sense … All of which function perfectly with the exception of these dang plugs and only with Sense’s local networking (Cloud functionality is always good).

I’m willing to experiment if someone can help me set up some sort of monitoring. I know EXACTLY when it will next occur. This is a bit beyond my normal skill set.

I think we’ve clearly established that there is a defect with the plugs. No question of that. But the problem could be the result of an external event, or of an internal issue. I don’t think there’s enough to say which, although I would lean toward external event because 1) it happens to some but not to others; and 2) it happens to all your plugs at the same time.

Agreed that DHCP renewal can be ruled out.

I would still take the step of setting one or more of the plugs to be local only. Just to cut down the number of variables. Unless you really need remote access to turn on/off the plug, it can’t hurt. And if it does prevent the issue, that would be very telling.

Moving the plugs and Sense to an isolated VLAN might be beneficial, but this requires networking expertise. Getting a packet capture of the plugs leading up to the event would be of great value, but this requires even more network expertise…

OK, I just moved two of the five plugs to local only. We’ll see what happens Friday night at 10;50PM.

You might have stated above but this has gotten so long I don’t have time to reread everything over the next 30 minutes.

  1. Router for DHCP? Model?

  2. Total Clients on Network?

  3. WiFi - Access Points or what? Make/ Model? If multiple are the all wired to router or mesh or sorting like AC line linked?

What I’m trying to get at is what level of equipment are you using in your network, especially as a number of routers are for a small number of clients and limited bandwidth/processing within.

As opposed to Commercial Quality systems (which very few will have in their resident).

This might not make a bit of difference, but it would be good to list what your system is capable of.

As a side note, the KS115 on my garage freezer went bonkers exactly 5 days after I installed last Tuesday. Was flashing Blue and the temp went from -10F to 23F before I caught it by chance.

When yours are locking, are you still seeing a solid blue?

That’s all been covered ad nauseum.:slight_smile: The problem has existed through a variety of equipment, the only constant is a Verizon FIOS Quantum gateway router G1100.

  1. DHCP originally from G1100 router, Current DHCP setup: Apple Xserve running macOS Server v10.11.6
  2. 81
  3. APs: Four different configurations during this time period.
    a) Three Apple Airport Extremes: All Ethernet home runs to main switch (Netgear GS748T 48 port switch).
    b) 2 Ubiquity U6-Lite’s ethernet home runs to main switch
    c) Current setup: Netgear Orbi Tri-Band Mesh WiFi 6 System, mesh, one main unit ethernet to switch, two satellites on dedicated 5GH backhaul.
    d) for troubleshooting only, the Sense and the five plugs were briefly put on the FIOS router’s 2.4GHZ wifi which is normally and currently off.
  4. Yes the Kasa lights are still blue and still function through the Kasa cloud, just don’t respond locally to the Sense.

Other than the foray into briefly putting the Sense/KP115’s on the FIOS’s 2.4GHZ, the other WIFI changes were to chase a system that reliably held onto the 81 devices (the bulk of which are 2.4GHZ IoT lights, switches etc. The Apple Airports served me well for many years until I got into a large number of IoT devices, then they had issues holding on to that many clients. The Ubiquity AP’s did a good job with that but the 5GHz coverage of the two available units was subpar compared to the Apple’s. The Netgear Orbi mesh has been fantastic. No issues at all with 81 devices, fantastic speeds and coverage throughout the house. The only thing is this Sense/KP115 issue which has persisted through all the WIFI and DHCP server changes. EVERYTHING else is as close to perfect as I could hope for. I would replace the FIOS router in a heartbeat but since I also have FIOS TV and need the MoCA connections for the boxes…

You can get a new router. Run Cat 5e or 6 from ONT to Router.

Have techs at FiOS change in ONT setting.

Run a patch cord into your FiOS router from the Router.

Turn off WiFi, DHCP et all in FiOS Router.

Also, you thought about running 2 networks - one for each band?

Turn off 5G on Ubiqiti and 2G on the Orbi units. Run a different system for each band, using the best of each.

The router is already etherneted to the ONT. I have gigabit up and down. The Ubiquiti’s have been sold. I’d rather chuck the Sense and the KP115’s in the trash can then go down another WIFI setup change route LOL. The Orbi is working great with the IoT 2.4 GHz and 5GHz devices. I have extreme doubts that another WIFI change would make any difference for these buggy kp115’s.

Believe me. You’ll find some serious questions from me about the KASA units on some threads over the past 2 weeks. So I’m not dismissing anything.

But in some other instances on some various friends residential systems, I have seen issues putting 50+ clients on a residential network. They just were not speced for that. The older and cheaper, the worse it gets.

Not saying that’s the case for you, but I have seen issues on some other residential units with heavy client loads.

I would also point out as the Sense unit queries all KASA units every x seconds, so the 5 days would be close to the same amount of queries every 5 days if this where the system overloads and gives up, be it cache not flushing properly or whatever.

Again, just thinking off the wall and brainstorming possibilities.

Here’s one for you. I have devices disappearing in the KASA app but not the Sense App.

And one in KASA app not appearing in Sense App any longer.

4 devices
IT RACK - Peplink 710 Router
IT RACK - Dell Ethernet x1052p
IT RACK - Wow Cable Modem
IT RACK - FiOS WiFi Router

All were showing on both Apps several days ago.


That’s exactly why I went down the rabbit hole on WIFI AP’s - unrelated to this Kasa issue, The Apple AP’s just couldn’t handle the load once I started adding IoT devices. The Ubilquity AP’s were fine in that regard but couldn’t cover the entire house. The Netgear Orbi system is neither old nor cheap. It works flawlessly with 81 devices. I haven’t had a light switch wig out once since putting it in and I’m getting 500+ mb/s on WIFI 6 devices everywhere in the house. Now these KP115’s/Sense - Ugh.

Just one caution @Beachcomber - Even if a smart plug shows as visible and on in the top level Kasa app, that doesn’t mean it is responding. When my HS110s turned into non-responsive Zombies, they looked good at the top level in the Kasa app. But when I pushed down into a he Energy display, this:

2 Likes

I got two KP115s yesterday and installed them on my server rack UPSes (one KP115 on the main plug for each UPS). They are installed right next to each other on the same outlet. One of them has been working flawlessly. The other has random dropouts (both in Sense and in the Kasa app). I am going to try some things to see if it’s a network interference issue from having them right next to each other, but I was curious: does anyone else have two of these installed on the same outlet? I know this isn’t a Sense issue (because it goes offline in the Kasa app too), but there may be more KP115 experience here than at TP-Link support, so I thought I’d ask here. I didn’t see anything on the TP-Link website that discouraged use of these side-by-side.

I just passed 5 days from my previous HS110 anomaly - 5 days, 11 hours right now. So far no n/a’s, though the Sense app response from the mother ship appears to be running a bit slow.

We just resolved this on our end, you should see response rate decreasing shortly.

2 Likes