Securing and stabilizing the Vera by taking it off the grid

[quote=“rafale77, post:140, topic:199140”]Thanks amg0. When you mentioned newer firmware, I suppose you are referring to the vera firmware because all other controllers I have (ZWay and home assistant) both still control the nightly heal regardless of the zwave firmware.

I am fairly certain that I have not had any nightly heal in weeks. It used to be very noticeable when it was healing as I had crazy lags at hours I am still awake. I have triggered a manual heal in the past on the vera which left me with 5 devices showing undetected (while still working perfectly) so I am really not keen on doing a vera heal.[/quote]

I would agree. from the back of my memory I think they were talking the zwave chip firmware ( which some VERA firmware did update at some point )

Just had a Luup reload caused by the zwave command queue being overwhelmed so this confirms that there is a weakness in handling a long list of zwave commands. I just don’t know what the limit is…

I determined that the variable zwave_heal, reported by ALTUI under the Table Controllers menu, really seems to indicate a Z-wave save. It will update anytime the Z-Wave network is backed up. Therefore, I’m not sure when the automatic heal runs.

There is a “Z-Wave Heal” button in ALTUI that generates this output in the blue banner on the test system:

Heal Stage : SUCCESS! Z-Wave network heal successful
Heal : Node 7 0 good routes out of 0
ZWave : Configuring Z-Wave devices in your system.

The zwave_heal timestamp did not update.

I’m not sure what all this means.

That’s interesting as I am not seeing this at all. My zwave network is being backed up by the vera daily yet ALTUI’s last heal date is from over a month ago.

[quote=“HSD99, post:143, topic:199140”]There is a “Z-Wave Heal” button in ALTUI that generates this output in the blue banner on the test system:

Heal Stage : SUCCESS! Z-Wave network heal successful
Heal : Node 7 0 good routes out of 0
ZWave : Configuring Z-Wave devices in your system.

The zwave_heal timestamp did not update.

I’m not sure what all this means. [/quote]

I believe this only checks whether the device is reachable and updates neighbor nodes.

Also I often have luup reloads due to command queue issues when I have a device go out of battery and becoming unresponsive. There is some work needed to in the luup engine to handle situations with no ack better.

I’m running a slightly modified version of the daily backup script you kindly supplied a while back. The zwave_heal timestamp updates every night when the backup runs on both Vera Plus hubs.

Perhaps try a manual Z-wavre backup and see if the timestamp changes?

I just did and it did not update the timestamp. My previous experience with it also never had the time stamp correspond to a backup time. It was always some random time between midnight and 3am. It is currently stuck in October.

Stranger and stranger…

maybe amg0 can chime in? I am also wondering if it is the time when the zwave dongle is backed up and sent to the mios server. Since I killed the tunnel to mios it can’t send anything there anymore.
@HSD99 did you apply all the mods I made? Maybe you are still connected to the server?

@rafale77, my test unit has the mods in your script applied to it. My production system is still on the grid. They both perform the same way as far as the Z-Wave back-up/zwave_heal timestamp is concerned.

The Mios gateway sees the test unit as being off-line.

And you are on the latest UI7 firmware and ALTUI version? (I doubt the ALTUI version does anything but just in case). This is quite strange. I am on the 1.7.3831 which is the one before last. I have a test unit on 1.7.4001 on which I just tried to run a manual zwave backup and also did not update the zwave_heal timestamp. Actually starting a manual heal on the test unit doesn’t update the timestamp either… I have no zwave device on that unit so I am not sure what is happening with it.

Both units are on 1.7.4001 with the latest ALTUI. The test unit only has 3 Z-Wave devices. Neither unit controls Zigbee devices. The production machine is extrooted.

[quote=“rafale77, post:133, topic:199140”]Going on 9th day without a Luup reload. I think the only unwanted luup reloads left are the time change and 1st of the month and zwave related ones.

And oddly I am also not seeing any nightly heal for the past month which is making my zwave network very happy and stable. I still don’t know how I disabled it though.[/quote]

Rafale, how are you monitoring luup reloads?

[quote=“Sender, post:152, topic:199140”][quote=“rafale77, post:133, topic:199140”]Going on 9th day without a Luup reload. I think the only unwanted luup reloads left are the time change and 1st of the month and zwave related ones.

And oddly I am also not seeing any nightly heal for the past month which is making my zwave network very happy and stable. I still don’t know how I disabled it though.[/quote]

Rafale, how are you monitoring luup reloads?[/quote]

I have a notification setup at the end my startup lua. Everytime my startup lua is run, it sends a notification through pushover so I get an alarm on my phone. I used to use the vera push notification service but the mios server was too unreliable, sporadically having a lot of delays or even being down so I switched to pushover on home assistant. To the vera it is just an http post call to my home assistant bridge.

FWIW, I monitor the uptime of my Veras (as tracked by EventWatcher and archived in openLuup’s Data Historian.)

Compare and contrast the current uptimes of my three Veras and two openLuup installations!

Very Nice AK. I think I will end up setting something like this up at some point.

My last Luup reload as of now is the dreaded midnight on the first of the month (PST) so it?s been over 7 days. The last unexpected reload was seemingly after I had scenes overload the zwave command queue due to a device having ran out of battery and therefore not responding to a command and clogging the command queue. I am hoping to be to hit a month. Let?s see how this goes.

Sigh. Had to do a couple of luup reloads:

  1. The vera somehow completely lost control of zwave after I got home one night. The house mode change causes a lot of zwave commands and the queue appears to have been too much to handle this time because it was combined with one of my ecovent having run out of battery. The vents and my Yale locks are battery operated but are considered always on devices so they are relays in the network mesh. I ended up deciding to manually reload luup and got everything working again before changing batteries on the vent.
  2. Another of my vents ran out of battery and this time the vera reloaded luup on it’s own after 10min of inability to send any zwave commands.

During a luup reload, the zwave API library does not appear to be reloaded so the command queue is definitely in the luup engine and seems poorly written to handle large queues. Having observed how Zway handles it, the vera has a lot to catch up to.

…I didn’t think that you were a fan of ZWay any more?

…I didn’t think that you were a fan of ZWay any more?[/quote]

Got me on that one. Zway is definitely not my bridge and I am not actively using it. I have just been using it as a tool to compare.
Lots of pros and cons to balance for a device bridge.

ZWay:
Pros: Very good/transparent zwave API/queue handling, supported by openLuup
Cons: slow support when it exists. No support for secure class key input. All kinds of firmware upgrade problems.

Hubitat:
Pros: Device support (inherited mostly from SmartThing). Good support and stability.
Cons: No official local API so it is difficult to control from another controller. Designed to be a standalone controller.

Vera:
Pros: Zwave and Zigbee(plus and secure) device support. support secure class key input.
Cons: underpowered Hardware for the software it runs. Obsolete OS. Poor Stability and zwave command queuing.

Homeassistant/openZwave:
Pros: With the combo stick, supports zwave and signee. Free. Very well supported by community. Supports secure key input.
Cons: Less readable command queue compared to Zway. No openLuup plugin. Device support is WIP.

I have had an interesting experience over the last couple of years with vera. Except for some exceptions, once I isolated Vera to its own subnet/vlan on my network it can go several weeks without an unplanned LUUP reload. I get notifications via Vera Alerts when a reload occurs. This is with about 60 devices, at least a dozen plugins (including cloud dependent), Hass integration, and several Imperihome tablets always polling. I am (and usually in the past have) running the latest on my Vera Plus.

I do have others issues with some Zwave devices dropping and some logging issues I encountered, however from the constant LUUP restarts, isolating it to its own quiet subnet has solved the problem for me. If any of you just want to try stabilizing your Vera, I recommend trying it.

Hi rafale77, how do you view or leverage the z-wave queue you talked about several times in this thread? I use Elvira the excel file to view the logs which is very helpful but I’m not sure where else to look to see something like this.