Hi guys.
For 3 times I have this recurrent behavior in my RPI3+.
My internet connection gets “stuck”:
As soon as I arrive home, I access locally to PI SNO, stop the container, “rm” it, start again…
And poof…everything is working as it was, smoothly!
Anyone experienced this? Does it even make sense?
Is it possible its doing a garbage collect and lagging the rpi to a point where its unable to do anything else? Also do you remember when the first time happened was it recently or is it random occurrences?
No, it was only recently, it started about 2 weeks ago!
I even limited concurrent connections to 2, but only minor improvements.
If I stop the container, voila, everything works properly.
What about when you just stop the node? Why do you have to remove the container?.
Yeah, that´s what I´m saying, if i just stop the container everything works, WhatsApp video calls included, perfect!
Now that I think about it, yes, issue started more less at the same time as garbage collection started.
Is there any way to restrain this feature?
It´s really constraining to my girlfriend´s work.
If not, I´d have to consider disable this node.
I think theres a way to disable it in config but I havent seen anything in the config to do so, As I noticed when garbage collection is running for about 8 hours my node is very sluggish. I dont think a RPI 3 has enough power to power though this.
can’t really be a problem with storj. I assume you have read the requirements of running a node - in terms of bandwidth available and that you networking will handle everything? Is the RPI new?
Just saying, I´m an IT and Network admin and this is my 5th node 
My other 2 Windows nodes are running great, no issues.
I think there maybe an underlying issue with garbage collection running though because while its running your still trying to get and send data at the sametime. Its not just you I think ud notice it less if you were running in a raid or not running a mechanical drive. But even on my windows node I have notice my mechanical drive at 100% for a long time.
Completely agree with you. In addtion, I´ve notice that it starts at 13:00h GMT+1 approximately.
Yeah but it seems to run for a long time though and it shouldnt happen everyday and ive seen it run every day so far.
Any how we can control this, or is the team working on a fix?
This should be restrained to x% of system resources occupation.
I have mentioned it but as of right now I think its slowly becoming more of an issue, Even more so when your node has alot of data its almost like its just recopying the same data everyday.
It´s my case…node is 1.6TB out of 1.8!
Or at least this should be configurable by time. I wouldn´t mind it ran through the night for instance.
I would hope in production it would only run when network is slow. Because if network is being heavily used there no way my node is going to be able to keep up while running garbage collection.
Or by time. Network slow is not acceptable for me.
It should be configurable by time or system resources.
Maybe by time but I dont know if by system resource is going to be possible cause it doesnt read system resources.
Engineers should really have a look at it.
i think your node i lagging now because of bit egress going now. Becase garbage collector cant efect other PCs on network, or your Girlfriend use same RPI to work? check your router for statistics of bandwidth your network just can be at limit, thats all. I had to take 300/300 ethernet to meet my nodes Egress needs.
This sounds like an issue with your modem/router. Unless you are routing all of your internet traffic through your raspberry pi. I could also see this happening with just DNS traffic like with a piHole setup.