Ever since I upgraded to the Zooz 800 Series stick my network performance has been slow but then sometimes reliable. Not sure if there is too much network traffic happening? I have 36 devices and everything has been placed in a position that works fairly reliably but there is still a lot of slowness.
Any chance someone can read my log and see if there is something going on that I can’t triage myself?
Hi,
I’m not going to wade though a log (sorry), but from past experience, the typical issue with Z-Wave delays is attempting to communicate with a dead / removed device that has not been excluded from the controller.
I used to see random ~2min delays after a device died and was replaced. The coordinator kept trying the dead device as it had not bee excluded from the network.
Another similar issue is a comms dead spot, perhaps due to interference. Adding more mains-powered devices is the typical fix.
A network heal can help, but only if ALL nodes are healthy and available (even after a battery powered device sleep) to respond.
A firmware update might force a heal, and make the coordinator keep trying a long dead node - that has happened. Once, I had to use the manufacturer’s software to remove a node as other methods failed.
I have a lot more devices than 36, so that should not be an issue.
The logs show problems with being able to communicate well to devices. They do not show a ton of traffic so it does not appear to be a high data volume issue.
It looks like this node went dead and then was pinged and it took 8 routing attempts to reach it. Are you using an automation to automatically ping dead nodes? If you you may want to disable that to figure out which nodes are problematic. Have you manually configured priority routes? If so I’d delete those as you should not need them. So it looks like a weak mesh. I’d look at the stick placement, extension cable, make sure it’s away from WIFI, computers and materials like brick or stone. In the network graph look at the part to node 33 - does it make sense? You can also heal one by one the line powered devices, starting from the closest. What exactly is mode 33
2024-01-10T01:32:42.146Z CNTRLR » [Node 033] pinging the node...
2024-01-10T01:32:42.153Z DRIVER all queues busy
2024-01-10T01:32:42.157Z SERIAL » 0x010f00a90001002101002500000000cb96 (17 bytes)
2024-01-10T01:32:42.158Z DRIVER » [Node 033] [REQ] [SendDataBridge]
│ source node id: 1
│ transmit options: 0x25
│ callback id: 203
└─[NoOperationCC]
2024-01-10T01:32:42.161Z SERIAL « [ACK] (0x06)
2024-01-10T01:32:42.164Z SERIAL « 0x010401a90152 (6 bytes)
2024-01-10T01:32:42.166Z SERIAL » [ACK] (0x06)
2024-01-10T01:32:42.168Z DRIVER « [RES] [SendDataBridge]
was sent: true
2024-01-10T01:32:46.993Z SERIAL « 0x011d00a9cb0001e202b27f7f7f7f01010503290000020800007f7f7f7f7f89 (31 bytes)
2024-01-10T01:32:46.996Z SERIAL » [ACK] (0x06)
2024-01-10T01:32:46.999Z DRIVER « [REQ] [SendDataBridge]
callback id: 203
transmit status: OK, took 4820 ms
repeater node IDs: 3, 41
routing attempts: 8
protocol & route speed: Z-Wave, 40 kbit/s
routing scheme: Auto Route
ACK RSSI: -78 dBm
ACK RSSI on repeaters: N/A, N/A
ACK channel no.: 1
TX channel no.: 1
2024-01-10T01:32:47.011Z CNTRLR [Node 033] The node is now alive.
2024-01-10T01:32:47.076Z CNTRLR Retrieving priority route to node 33...
I think I fixed Node 33 as it’s responding now. (Not sure why it died). This node is right next to the 800 stick. It’s a siren alarm no security
I do have a pinging of dead nodes, that I might want to look at it.
I didn’t realize my mesh is so weak, it’s possible that I have a ton of interference from the cabinet that I leave it in. Maybe I need a longer extension cord? I know I had issue when I had it on the second floor and moving it to the first floor where I had more devices was better.
I did set a bunch of primary routes because I thought the mesh didn’t look or work right. I’ll go ahead and delete that based on your recommendation
I have 4 floors in the house but I have devices on every floor
I’m going to make the changes you suggest and see what happens and then post the logs again. Thank you for the suggestions!
I’m not doing inclusion, although I saw a node that has smart start enabled but zwave doesn’t realize that it’s that node. I might need to readd the device to fix this as well. It’s definitely giving me good ideas to fix it.
Just an idea if that works for you. My wifi network is a Google Nest mesh and I connected my HA Yellow Box to a secondary pod, away from all my other systems. This works really well. BTW I mainly did this because my cable (Comcast) comes in at the far end of my home and by connecting to the secondary wifi pod, my Yellow Box is in the center of the home.
That’s a great idea, I did try that, but then I was having issues on the first floor. Although I had it in a box and maybe not using an extension cord. Maybe I’ll do that and move it with a large extension cord. Thanks!
Got an extension cord and it is outside the box of electronics
I reincluded a smart switch that I think was trying to smart pair, even though it was already connected to my network.
I rebuilt routes starting with the closest devices
Is there anything else that looks weird in the logs? I did a few test to see the connections on zwave and did not get great scores even after all the changes that were made
Node 39 is very active, I don’t know what the SecurityCCNonceGet is doing but it is happening a lot and it probably consuming a lot of your zwave bandwidth. I don’t have security on my network and this is an area I don’t know much about.
Node 30 is also very active, sometimes is sending alarm reports every 10ms.
Wow this is super helpful! I’m wondering if I should just reconnect everything without S2 security. I have some devices that don’t have security so not sure how important it is unless everything is on it.
All of those active nodes are my motion sensors so that makes sense that it is most of my activity. I’m not sure how to disable alarm reports, that’s really interesting.
Secure devices cannot communicate through insecure devices so this may be weakening your overall mesh and causing issues. From what I’ve read hear you are either all in with everything being on S2 or all out.
Look at the configuration parameters for those motion sensors and/or contact the vendor to ask them what is going on.
Is any device plugged in considered a repeater? I have a bunch of switches that are connected via S2, some sensors S0 and other devices that are connect insecurely as they don’t support secure connections
Are all of your devices connected without security? I’m wondering if I should skip out on the security as I already have some devices that aren’t secure.
I have 3 light switches in our master bathroom and I was to do the association with 2 of the 3, as 1 is secured S2 and the other 2 are not. Next project is to unsecure the 1 switch.
I would start by excluding Node 39 and re-include without S0. As previously mentioned it’s spamming your network because of encryption problems. It also has the slowest link possible (9.6 kbit/s), and then at the end of the log it starts sending unencrypted reports anyways (which I’m surprised ZJS is allowing).
Okay I redid all the S0 motion sensors except my august lock that isn’t having any issues right now so I left that as S0. Really appreciate everyone’s advice.
Here is the latest snippet of logs. I think it’s better? Also my network graph doesn’t look great even after healing nodes. I’m wondering if I should move the network to the second floor as I have 4 floors in the house the the controller is on the first floor. Still not sure why the backdoor isn’t connecting directly to the hub.
this post!