@girish ooooo - so each backup stands on its own. That makes sense from a recovery perspective. I suspect it'd be a groovie feature to add though for Nextclouders and Jellyfin types so they know they have one good backup. In my case, disable backups of Jellyfin and back up and goin.
@girish It's no problem - just a tad bit of overhead and I use the S3 for other backups anyways.
Yeah, the volumes work great - in my instance it looks like this:
-External hard drive via USB3 hooked to physical box
-Hard drive volume created on physical box with an 8 TB VHDX on there.
-8TB VHDX passed through to Cloudron VM
-Cloudron VM mounts it as an xfs target
-Pass mount (the /mnt requirement tripped me up but is smart to require it) to Cloudron
-Mount volume into Minio container (/media/backuptarget)
-Reconfigure minio to default to /media/backuptarget instead of /app/data/data
(The last bit is why minio CLI's fuss cause /app/data/data is hardcoded in a few places).
@girish Indeed - definitely Go, not seeing anything in the logs other than say a site not resolving which causes the trigger to fire - nothing to console about the message going out - maybe I can find a debug=true string somewhere.
I believe @Lonk 's VPN project may be helpful here as VPNs track their packets much more so than general network interfaces which could provide for greater visibility for any app by simply adding it to the VPN profile & data path.
Until @girish gets a more general solution in for all of Cloudron and perhaps containers in general.
@msbt I have upped the memory to 2GB and since, this error : Oct 08 15:22:37 /app/pkg/start.sh: line 82: 101 Killed gosu cloudron:cloudron python3 -m synapse.app.homeserver --config-path /app/data/configs/homeserver.yaml
is completely gone.
But I'm still stuck at the same : even joining little rooms does not work, or works "randomly" ?