My router (Flint 2) becomes unresponsive after an application consumes a lot of RAM
Unresponsive means no response from the router at all: unable to connect to WiFi (both 2.5GHz and 5GHz) for old and new clients, unable to connect to the web interface through cable or GLiNet application, internet stops working, ... etc.
I am running syncthing on the router to backup folders to my hard disk connected to the USB port
97% of RAM is used (shown in the "Overview" panel in web UI and GLiNet application)
syncthing runs as root user. This is advised against but it's necessary for syncthing to access mounted hard drives
AdGuard runs with large lists
Symptoms
Mobile devices only continue working on WiFi (does not include laptops or PCs). After a while internet gets cut off and I am unable to connect to WiFi interface (2.5GHz or 5GHz)
Router UI is not accessible via web interface or GLiNet application
Question
I am assuming critical router subsystems gets killed because of syncthing ? I am unable to provide a log as I needed to power off and on the router to get it working, with no logs in the "Crash Log" tab, only logs for unexecpted filesystem shutdown and that I should run fsck
Where do you thing I can start debugging this ? It happens once every few hours
Did you start to see this problem when you added blocklists to Adguard Home?
If so, disable Adguard Home then, remove the lists that you recently added.
I started to see the problem after adding more folders to synchronize in syncthing, but AdGuard might be the culprit. I will check the link you posted. Thanks
I'm not familiar with Syncthing and may be totally misunderstanding the setup, but my initial thought is that the router is not the best host from which to run this kind of application... It seems like it would be preferrable to have the sync app run on the pc and automatically begin copying files when it detects the network drive is available.
So far, disabling a few lists in Adgurad and limiting Adguard's Go stack memory as @Renato shared seems to do the trick. For now, everything is stable. Thanks all for your input