i seem to be seeing about a 25kb memory leak after each firmware backup on my hubs.. think that is whay eventually it gets low enough and i need to reboot. For now i am going to disable the nightly backup via my nas.
ie
you can see the cpu spike at the end of the graph during the backup and the loss in memory.
Today's upgrade addressed a possible memory leak is it still happening after that?
yes unfortunately that last event in the post was on the latest build you can see the big spiker where it updated.. i regularly do a backup immediately ie 30 minutes after an upgrade so i have one
Does a hub reboot fix the memory leak?
Here is a second data.point.
My.other hub .. see the drop in.memory every nighy after thr maint cycle where a backup.occurs.
I'm seeing the same thing on my C-5. Once it recovered after the drop, but usually it does not. I see an increase in DB size. Sorry I don't have a pretty chart. Not sure how to create one.
Not sure why this is being ignored
I can get about 4-5 days before memory drops below 170 or so and i have to to reboot.
Yup seeing it here on my C7 too.
I'm seeing mem issues as well at times. I'm fairly new to HE after migrating 100+ devices to Hubi in January (finally!).
The experience has been mostly great but the occasional glitch seems to coincide with free mem dropping to around 250 kb. At least that's where it was when my z-wave radio stopped responding last week. That issue only happened once randomly, but I set a notification for when free mem drops below 275k anyway. So I can at least monitor logs or reboot.
I've removed or cleaned up every App I use a week ago. Last week's free mem decline looked a bit better. I notice today's hub update includes a mention of GC improvement. Maybe that will help.
Guys the drop is related to the backup but the recovery is related to how Garbage collection works. It won't immediately drop something out of memory, It is more likely it is remaining in memory for some other reason GC has determined it might be useful. especially if it doesn't have memory demain at the time requiring to be moved to a lower priority memory location.
The below graphic also shows the memory usage and recovery from the Backup Process. You can see dips at around 2:15 several mornings and then another at 4:40am. One of them is the nightly scheduled job on the hub and the other is created by my unraid server running a simply script to create and download a new backup. The key with this graph is you can see that it is already down to around or even below 200MB of memory. The hub is was running fine through this interval, but one key factor is that towards the end of that graph the memory drop was practically stable and not droping much at all any more.
Memory consumption is based on demand so you have to ask yourself what are you doing that is continue to pull that much memory. but i really don't think memory loss is being caused by the backup procedure.
What version are you all on? Iām on 2.3.4.123 on my C7 and memory got really low recently and it rebooted 
.
| Last Restart | 2023-02-19 12:19:41 |
|---|---|
| Uptime | 1d,10h,30m,0s |
Long term history
Last day or so:
Not entirely sure this has been resolved. If I have daily backups off, my hub runs consistently around 320MB of free memory (first screenshot). With daily backups set at 3am, over the course of 5 days my free memory drops by at least 100MB (additional screenshots taken after 1/3/5 days with daily backups on).
I migrated from the C7 to the C8 yesterday afternoon. For the first time that I'm aware of I didn't experience the usual post reboot overnight 50 to 100K drop at the first nightly cleanup.
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.
Download the Hubitat app
