I am on UI7 (1.7.569), have 155 devices (of which 5 are camera’s) in 16 rooms with 46 scenes, controlled by the following 19 plugins (Kira is not listed):
Is this too much? If not what is the limit and if it is, what is the best and easiest way to upgrade.
I am willing to buy extra hardware or even make use of my old Vera 2 if that makes any sense.
When I ssh/top into my unit it shows 173% of memory use:
Mem: 109696K used, 17760K free, 0K shrd, 8624K buff, 41252K cached
CPU: 14% usr 11% sys 0% nic 68% idle 0% io 0% irq 3% sirq
Load average: 0.98 1.13 1.03 1/136 2912
PID PPID USER STAT VSZ %MEM %CPU COMMAND
31142 2229 root S 215m 173% 14% /usr/bin/LuaUPnP
2128 1 root S 748 1% 2% /usr/bin/luci-bwc -d
5155 2450 root S 6932 5% 1% /usr/bin/NetworkMonitor
4144 1 root S 3924 3% 0% lighttpd -f /etc/lighttpd.conf
2436 2425 root R 1680 1% 0% top
544 1 root S 1928 2% 0% syslogd -C256
2455 1 root S 1864 1% 0% /bin/sh //usr/bin/cmh-ra-daemon.sh 12
2005 1 root S 1696 1% 0% crond -c /etc/crontabs -l 5
2229 1 root S 1688 1% 0% /bin/sh /usr/bin/Start_LuaUPnP.sh
1 0 root S 1684 1% 0% init
2425 2300 root S 1684 1% 0% -ash
514 1 root S 1684 1% 0% init
2159 1 root S 1680 1% 0% watchdog -t 5 /dev/watchdog
2450 1 root S 1680 1% 0% /bin/sh /usr/bin/Start_NetworkMonitor
2252 2229 root S 1676 1% 0% /bin/sh /usr/bin/Start_serproxy.sh
2845 31142 root S 1676 1% 0% /bin/sh -c ping -c 1 192.168.1.202
2844 31142 root S 1676 1% 0% /bin/sh -c ping -c 1 192.168.1.203
31755 31718 root S 1676 1% 0% /bin/sh -c ser2net -n -C "3483:raw:0
546 1 root S 1672 1% 0% klogd
2846 2844 root S 1672 1% 0% ping -c 1 192.168.1.203
I do have very frequent luup restarts, over 20 a day, sometimes within minutes sometimes within hours and every now and then locking up the Vera, needing a power cycle to get back on its feet. MVC has looked into it numerous times but there only advice is to remove plugin’s.
However, if you wanted to offload some stuff, you could at least move DataYours to your other Vera, (or indeed a RPi, or BBB or almost any other Unix machine) leaving only DataWatcher collecting the data on Vera 3.
I’m currently working on a Vera emulation environment (calling it “openLuup”) to run non-Zwave plugins and virtual devices unmodified. I’ve posted a version which will do the DataYours thing above, but internally I also have EventWatcher, Netatmo, ALTUI, RBLuaTest also running fine… I simply didn’t want to throw away all the Vera plugins and, from a development point of view, couldn’t bear to move back to Java from Lua.
The [tt]openLuup[/tt] environment is written in pure Lua, with a single external dependence being the LuaSocket library - both usually very easy to install (and very small). It implements its own HTTP server on port 3480, but relies on an existing port 80 server (which I assume the NAS has already.) Aside from that, you need to be able to create a directory tree for the vital bits of Vera, specifically, [tt]/www/cmh/skins/default/icons[/tt] (for UI5 emulation) and [tt]/etc/cmh-ludl/[/tt] (for all the device XML and Lua files.)
You have not indicated if you have more than one of the plugins … Total count is a factor!
How many Virtual Switches ?
How many Countdown Timers ?
How many Combination Switches ?
How many Ping Sensors ?
How many RGB controllers ?
…
Adding Z-Wave devices is relatively cheap … the memory is consumed by the controlling device … i.e. the internal Z-Wave controller.
The memory size is only a problem when it hits! There are periodic things that happen on Vera that grab memory for a short period of time.
And if Vera needs the memory at the same time … it will cause a restart.
But the fact that the LuaUPnP Virtual image size is 173% of memory indicates you can have a problem. The higher this number is above 100%
the more likely you can have a problem. This includes stack sizes for stacks that are never/less frequently used and some double counting.
The more often the configuration of events happens to hit some of these less frequently used stacks … and then you will have problems.
If that happens while some other periodic tasks on Vera are running … it will also happen sooner.
You might need to collapse # of plugins - Replace Virtual switch with a multi-switch,
Replace all of your countdown and combination plugins with a single PLEG/PLC combo.
I have seen Vera Restarts causes by the Ping Sensor … some times it takes forever for the ping to NOT respond and causes an overrun.
Best Home Automation shopping experience. Shop at Ezlo!