Not finding a general thread for active releases I am making a new bug report thread one as per instructions in the forums, and also tagging @gopher.ny
The current release, 2.3.6.144, seems to have a memory leak. After upgrading both my hubs they now run low on memory and ask for reboots. If neglected, the hubs will eventually freeze and become unresponsive. A power cycle fixes the problem, until memory runs low again.
I run the following additional apps:
HUB 1:
bleBox integration
Export/Import clone
Google Home
Groups and Scenes
Hubitat Dashboard
Mirror
Simple Automation Rules
HUB 2:
Basic rules
Chromecast Integration (beta)
Hubitat dashboard
Let's get started!
Rule Machine
Additional information:
Hub 1 has 4 custom drivers and 20 devices.
Hub 2 has 3 custom drivers and 21 devices (one virtual)
The only common drivers and code present on both hubs seems to be:
[Ikea TRADFRI Shortcut Button] (mast76)
[Konke ZigBee Motion Sensor] (muxa)
The only App that is using custom code is the bleBox integration, and it is only present on Hub 1 (so unlikely to be the problem as the issue is on both hubs).
Both hubs exhibit the same issue, though Hub 1 seems to run out of memory slightly faster.
I have never had memory issues before.
I am happy to upload log files or other information that may help in further investigations. I can also provide the Hub ID's in a PM.
The hubs have been rock-solid for 2+ years, so this is the first real issue I have experienced. That said, the WAF factor is obviously low, when I have to explain that I need to "reboot" the fan in the shower, the towel heater, and the motion sensors...
In the short term you may want to set up a rule in RM to reboot your hub automatically at 4am or similar, when memory gets <170,000 or so. That way your family (and thus you) will be somewhat protected. You can use the Hub Information Driver to track memory levels (among other things).
What version were you using before .144? Have you tried rolling back to the most recent release you were using before .144 to cofirm that the memory issue goes away then?
I am seeing a strange memory issue with a method I use in a lot of my drivers, not sure if it is related at all. Was testing a new driver and one user with a C5 kept getting out of memory errors always on the same line of code. It would cause the hub to get slow and miss zwave events until he rebooted.
void setParamStoredValue(Integer paramNum, Integer value) {
//Using Data (Map) instead of State Variables
TreeMap configsMap = getParamStoredMap()
configsMap[paramNum] = value
configsList[device.id][paramNum] = value
device.updateDataValue("configVals", configsMap.inspect())
}
The last line of the function with the inspect() is what the low memory errors were referencing. Its the only place I use that method in my drivers, not sure if that's the cause of it or not.
The C8 hub I was testing with saw a couple of large drops, one was 80Mb and hub is down to 190mb free now but still working.
Other than that, during beta testing and afterwards, memory has been very stable on both of my hubs.
As reported elsewhere here, I was having goofy memory problems with one hub until (at the excellent suggestion of a fellow user) I did a backup/soft reset/restore.
I would up doing the same with my other hubs as well, and experienced vastly increased available memory and a vastly decreased loss rate on all of them.
Yeah good point @SplatMan I would try doing a backup/restore. That fixed mine in the past but this most recent issue I found is different and maybe not related to OP issue.
I am having the same problem in one of my four C-7 hubs. After a reboot it shows 630 MB of free memory, but in as little as 3 hours it is down to 300 or less. I usually have to restart it every night to keep the hub running. By the end of the day it can get down to 170 or less.
Have you done a backup/restore? You just download a local backup and then restore it right away on that same settings page. This will also do a soft reset when you restore. For some reason this seems to fix a lot of memory issues people have been having with 2.3.6
The devs fixed something, and the issue is gone with the current firmware version. So all is good. I had a nightly reboot routine that kept things afloat, and I disabled it a week ago to see if the issue persisted. It did not.
I tried sifting through all the release notes (literally - took a while) but can't identify the item that fixed this.
In any case, the issue is solved for me, without doing the backup/reset/restore procedure.
I think it was 2.3.8 that had major memory work done during the beta.
We were posting charts every few days and they would make more tweaks to make it better.
