I’ve migrated up to a C8 for the latest radios. I’ve found they tend to be a bit more responsive then the C7 which was great but we began having complete outbound action failures on the zwave network. At some point every 24 hours, always between 7pm to 10pm at night the following issue starts.
I present a button to do a house action like turn on/off all lights in the kitchen and dining room to get ready for bed. Nothing happens.
I check the logs and see the button I pressed as received and recorded by the Hubitat. It also claims to have run the actions such as turn off all the lights
I check the Zwave logs and find no outbound commands sent to any of the zwave devices to turn off.
I run the following test:
I log into the Hubitat through my iphone app and attempt to turn on/off one light through the devices page.
I check the z wave logs and no message are sent to turn on / off a signal device.
What I’m seeing is the Hubitat stop sending Zwave commands to devices but will continue to receive them.
The only way can restore normal operation is a hard power cycle of the device. Standard reboots don’t appear to reset the radio. I even tried downgrading the firmware but that had no effect. The downgrade carried the radio issue through the reboot and again only a hard power cycle resolved the issue.
One other note, I did look to see if I had a regular event that correlated to the outage but couldn't find one.
I might be misunderstanding what you’re seeing, but IIRC the Z-wave log page is live only. Log entries don’t persist, so you’d need to have the page open while sending Z-wave commands to see anything.
Good question, I think you're asking if I'm watching the live z wave logs vs the past z wave logs. I'm checking the past z wave logs right after I attempt the action. In cases during the failure, no actions go out in the past logs. After I hard reboot, I do try the same action again and in those case I do see actions going out to lights in the past z wave logs. Does that make sense?
Does anyone know if there's a concept of a CPU load or resource load anywhere I can check? I have a relative big Z wave deployment of about 130 devices and I'm trying to see if something is consuming a lot of resources. I go through log device stats and app stats and they're outliers but I can't say if they're really causing any problems.
Sorry, you're right. the "past" logs I was looking at are the general logs. I do keep the zwave logs up also and don't see actions coming in as if something is taking up all the system resources. That's a guess. I don't have proof yet. I see the recommendation for the Hub information drive above. I just installed it and have the following cpu info.
cpu15Min : 0.25
cpu15Pct : 6.25
cpu5Min : 0.27
cpuPct : 6.75
freeMem15 : 417880
freeMemory : 410168
That doesn't feel bad to me. could be wrong. I must have something else going on. Need to dig more.
I couldn't find any issues in the logs or resource constraints that gave me any indication what's wrong. I decided to roll back to backup I used to perform the migration from the C7 to the C8 and wait a few days to see if the issues come back. Note, the only changes I've made in the C8 since I migrate were updates to the app "button controllers".
For now things have been stable. If things remain stable for the reminder of the week, I'm going to add the button controller configurations I had pervious to see if this is in fact an application issue not a z wave issue.
The reload of the first backup appeared to resolve the issue. Things have been stable for a week which is significant. At this point I'm going to consider this issue resolved. I wish I had better insight to why things were crashing.
