I am running OH 3.0 on Ubuntu 23.04. I am adding a lot of lifx bulb and now up to 57. Now when I go to add a new bulb is doesn’t bring up the screen to add a new device. It seems to have been slowing down as I add more devices.
I have configured my ubuntu with 4 cores and 32Gb of ram on a vmware virtual machine.
Is there a limit to the number of devices? is there an adjustment that I need to make?
In general no, and even if there were 57 wouldn’t be that many. However, there might be limitations with that specific binding.
What is your RAM and CPU usage? If there’s a memory leak or a process run amok that can cause slowdown.
Ultimately, if there is a bug here, you’ll need to update to the latest OH version to see if it’s still a problem. A lot of development has occurred in the 2+ years since OH 3.0 was released. OH 4.0 is about to be released in a couple of weeks.
the cpu and memory usage are minimal Im seeing no more than 4% on the cpu and about 2g on the memory. Also I put in the wrong version, im running 3.4.4.
out of the 57 devices so far about 53 of them are lifx bulbs
OK, following up with what @JustinG suggests, put the lifx binding into debug and see if you can see some delays there or something else fishy. For example, you physically command the light and it takes forever for the message to arrive in OH or the other way and commands from OH according to the log get sent right away but the light takes too long to change.
That can point to whether the problem is native to OH or outside of it.
the wifi network is a cisco 4800 enterprise AP. What is online is staying online. I have 1g connections from the AP back to the Cisco nexus switch. The openhab as mentioned is a virtual machine running on vmware. I did increase the number of cores from 4 to 8. This seems to have improved things a bit.
is there a recommendation document with scalability informatin like small medium and large deployments?
this is in a house, there are other wifi signals around but not many. I am getting vm lockups. im not sure if it the vm or vmware. Im going to move it to a completely different vmware instance. But what I see, the more devices, the sooner it locks up. It was once a day. now its every half hour and I have to reset the vm to recover.
No, but from a generic OH perspective, your’s is a small deployment. However, it might be large from a lifx binding perspective.
Given the nature of OH, Java, and modern operating systems, I struggle to see how OH could be the root cause of the VM locking up. It seems more likely that there is some other underlying problem (hardware?, hypervisor?) that OH is revealing simply by running. Adding more lights to OH is causing OH to poke at the problem more often perhaps. Maybe that might provide a clue.
Even when OH runs amok with memory leaks or bindings running out of control, at most it crashes OH, not the whole machine.
here’s an update on the problem. I have added more devices and am up to 73 lightbulb. what is happening now is that when i go to the things page to add a new device, it hangs. Usually if i restart the VM, the problem gets resolved. But now restarting the service or the vm itself, the things page still hangs.
I doubt this will be the same as your problem, but it may be of assistance - I’ve had a similar experience with periodic “slowdown”, although in my instance it was not correlated to more items / things.
In my case it wasn’t OH at all, but my fail2ban settings blocking unwanted connection attempts “forever” and then iptables would consume any and all available CPU. Can you nail down the process using top or similar (assuming this is Linux) to confirm that it’s OH-related?