zwaveCrashed Often (Every hour or so)

Had this issue on a c-7 as well, upgraded to a c-8 for the better antennas (and already see the advantages after a couple of hours), but getting zwaveCrashed every hour or so (Not exact, but on the c-7 it was often at :00 or :30 minutes after the hour).

When it happens, all the devices physically work but for 5-10 minutes the hub won't send or receive any messages from the z-wave network (so no automations work), and then it'll report zwaveCrashed, recover, then everything continues working.

Mostly just an annoyance, but a 5-10 minute outage for automations every hour or two (16 events yesterday on the c-7, 3 events so far today on the c-8 installed 3 hours ago) is having a pretty significant impact on WAF.

Submitted a support ticket too, but no place to put real information there, so I figured I'd start a topic here for public visibility and a better way to communicate as well

Debugging done so far (Mostly on the C-7, but checked on the C-8 as well):

  • No Z-wave ghosts: Removed ghosts (approx 3-4 mo ago) using a Silabs z-stick. None came back since, no help. Verified no ghosts on C-8 as well
  • Rebooted, and shutdown, remove power, wait 10 minutes, return power (Both C-7 and C-8)
  • Turned down reporting intervals on Inovelli devices. Switched 2 red-series devices to blue series (zigbee) to reduce network load
  • Soft reset hub (c-7), no change
  • Removed Aeotec home energy monitor (Associates S0 security, no way in hubitat to get it to associate security-free, lots of power reports)

I'm assuming at this point that it's some issue with the z-wave database or a rouge device, but I haven't been able to associate the failures with a specific device sending packets, or interacting with specific devices

ZWave details:

This sounds like a rogue device spamming the mesh...

2 Likes

I'd agree, but it's all pretty common devices, a few Jasco switches, some inoveli red series, some aeotec sensors and sirens, a single home seer fan switch, 3 kwikset locks (Ugh S0 security), and some Nortel (GoControl) sensors.

As fast as I know, everything I've got is pretty well behaved in general. I know the inoveli switches can be a little heavy on the power reporting by default, but there's very few events for their reporting watching the logs, maybe 3 or 4 an hour per switch

Anything I can do to try and figure out if it's a spammy device, and which one?

Use the Z-Wave stick you already have from removing ghosts, look at traffic with SiLabs PC Controller.

@zcorneli.iit

Can you PM me your hub id?

Sent.

Been working with @bobbyD from support, recommendation was to reset the Z-wave radio and rebuild the mesh (replacing devices as I go).

Haven't gotten many devices re-added yet, only about 4, going through it slowly today.

Got the rest of the mesh rebuilt, and for the first time in quite a while, no zwave crashes for over 24 hours.

2 Likes

Ya if it was a bad device you need to go through one by one and figure out which was not working or had a cluster of devices not working around.it.

Hard to.do.

Doesn't seem to be a bad device, since they're all readded and working, just a corrupt zwave radio database from early on in the c-7 era that I never fixed

2 Likes

Further update after a few days. No further crashes, and I even readded the dreaded Aeotec HEM gen5 on s0 security. Hubitat complains about it's excessive events, but the zwave network is rock solid and quick.

1 Like

You can use a stick to add that without s0

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.