New scenes not appearing in UI5 GUI

I have created a couple new scenes today, but they are not appearing the the UI5 browser GUI. I’ve closed my browser (chrome), saved, reloaded, rebooted, etc. I was able to see and copy in a scrolling browser log that a scene seems to have been created or is trying to be created but isn’t being successful. only one of the new scenes i’ve created shows up in the log when i search and the “code” around it looks like this:

“scenes”: {
“scenes_1000001”: {
“id”: 1000001,
“name”: “testscene”,
“room”: 2,
“groups”: [
{
“delay”: 0,
“actions”: [
{
“device”: “21”,
“service”: “urn:upnp-org:serviceId:SwitchPower1”,
“action”: “SetTarget”,
“arguments”: [
{
“name”: “newTargetValue”,
“value”: “1”

Does anyone have any ideas why this would be occurring or what I should be searching for in a log to indicate a system problem?

Thanks

I figured this out. I had recently installed the ‘Say The Weather’ plugin and that seemed to have caused the issue. Once I removed that and created another simple scene, it appears in the GUI. I know that the ‘Say The Weather’ plugin was designed for UI7, but the author had recently made some cosmetic workaround changes that made it usable within UI5. It seems some of the behind the scenes stuff for UI7 somehow broke my UI5. I’m running 1.5.672 so that a fibaro 4-1 motion works properly. I’ll post this over in the board for that plugin and if the author wants me to reinstall a newer version to test this out again I will.

This could also be a symptom of running out of memory space. If you are right on the edge, even a slight addition could prevent the configuration being saved when you restart. If the current configuration isn’t saved, you get the previous one - without your addition.

You may be able to free-up some memory if you can uninstall any non-essential plugins. This action does not remove the files from Vera so it sometimes helps to delete them manually using WinSCP. The files are in the folder /etc/cmh-ludl.

Thanks Rex

I installed eventwatcher and when I have a bit of free time I will grab some metrics from that for my system now without the Say The Weather plugin and then re-install it and see what it looks like afterwards.

I know you are a big advocate of PLEG. Does replacing normal Vera scenes with PLEG logic help reduce memory usage at all?

I know you are a big advocate of PLEG. Does replacing normal Vera scenes with PLEG logic help reduce memory usage at all?
NO.

@ctguess…thanks for taking a deeper dive into your system to see if you can determine a root cause.

@RexBeckett (or anyone else…) it is not clear to me how to exactly determine the “memory sate” of one’s unit. I’ve searched and seen references to the “top” command and the “free” command, but no real guidance on what to look for in terms of being an indicator that memory usage is getting at or near the “edge”. Any definitive guidance?

For example, here’s a snapshot of “top” for my own VeraLite UI7

Mem: 47324K used, 15156K free, 0K shrd, 0K buff, 18468K cached CPU: 46% usr 7% sys 0% nic 45% idle 0% io 0% irq 0% sirq Load average: 0.64 0.61 0.48 1/84 28198 PID PPID USER STAT VSZ %MEM %CPU COMMAND 3920 2790 root S 76860 123% 1% /usr/bin/LuaUPnP 3164 3065 root S 6892 11% 1% /usr/bin/NetworkMonitor 28010 28003 root R 1680 3% 0% top 5306 1 root S 1128 2% 0% /usr/sbin/dropbear -P /var/run/dropbe 2699 1 root S 748 1% 0% /usr/bin/luci-bwc -d 5206 1 root S 3928 6% 0% lighttpd -f /etc/lighttpd.conf 8052 7929 root S 3172 5% 0% -ash 2692 1 root S 2196 4% 0% /usr/bin/lua /tmp/upnp-event-proxy.lu 438 1 root S 1928 3% 0% syslogd -C256 3013 1 root S 1836 3% 0% /bin/sh //usr/bin/cmh-ra-daemon.sh 12 2618 1 root S 1696 3% 0% crond -c /etc/crontabs -l 5 2790 1 root S 1688 3% 0% /bin/sh /usr/bin/Start_LuaUPnP.sh 2516 1 root S 1688 3% 0% udhcpc -t 0 -i eth0:0 -b -p /var/run/ 1 0 root S 1684 3% 0% init 403 1 root S 1684 3% 0% init 28003 27877 root S 1684 3% 0% -ash 3065 1 root S 1680 3% 0% /bin/sh /usr/bin/Start_NetworkMonitor 2730 1 root S 1680 3% 0% watchdog -t 5 /dev/watchdog 2813 2790 root S 1676 3% 0% /bin/sh /usr/bin/Start_serproxy.sh 4409 4375 root S 1676 3% 0% /bin/sh -c ser2net -n -C "3481:raw:0

and here’s the results of “free”:

total used free shared buffers Mem: 62480 47252 15228 0 0 Swap: 0 0 0 Total: 62480 47252 15228

Any definitive guidance?

That one is tricky. The three commands most used are:

[ul][li]top[/li]
[li]free -h[/li]
[li]df -k[/li][/ul]

They will often give different answers for the amount of available memory and it isn’t necessarily helpful. Even when it appears there are thousands of free KB, a combination of concurrent activities/actions/requests can put the short-term requirements above the availability.

In your case I would be concerned that LuaUPnP could demand 123% of the available memory in some circumstances. It may not happen very often but it could be a cause of random restarts.

Thanks! I have removed a couple plug-ins from my own unit and even deleted the related plug-in files from /etc/cmh-ludl. Did not impact LuaUPnP memory. Is there anything in particular that impacts this? (Number of devices?)

Z-Wave devices have relatively little impact on memory usage. Plugins are the big eaters. LuaUPnP.log will also consume a cyclical amount.

You may not see the reduction in memory footprint unless you perform a full reboot or power-cycle.

I did actually perform a reboot (Settings → Net & WiFi -->Reboot vera) after I uninstalled and deleted my unused plugin files.

That being said, I do have a “large” number (24) of devices that are not z-wave, but are binary switches and dimmers that are children of a parent serial I/O device. I wouldn’t think that would take up much more memory than the parent device itself.

I wouldn't think that would take up much more memory than the parent device itself.

That was certainly the case on UI5. It is possible that this is different on UI7…

I just checked my Vera Lite. This is running 1.7.481. It has no Z-Wave devices and only four plugins. top shows LuUPnP is at 100%. It looks to me as though Vera Lite running UI7 does not have much room for plugins.

Thanks for your insight. I’m not too terribly concerned at the moment since It seems stable enough, I was just curious about “how things worked”.