I’m running 18 safenodes and their logs on a USB stick without any problems…any more and the internet router becomes strained.
safenode-manager add --count 18 --owner discordID --home-network --data-dir-path D:\ --log-dir-path D:\
6 Likes
I’ve had to settle on 10. 
That’s a long way from the 20 I thought I was going to be able to run on my 80/20 ADSL. 20 was causing me issues when I was working (yes, I still have to hold down a ‘square’ job). At 15 Netflix and Prime Video were flaky.
I’m considering running the 10 all the time and switching the extra 5 on for 18 hours a day when TV watching isn’t a thing.
3 Likes
Was the same for me when I was running 20 nodes from one machine (vdsl 50/20 with home network) but then I divided 20 nodes to two machines (10 nodes each) and now my netflix etc. at home is working with the nodes but for now where is not so much data on them, we’ll see in next days.
1 Like
I don’t know about the Launcher but it is available from safenode manager if you use the --version flag to add it:-
safenode-manager add --version 0.108.3-alpha.2 --home-network --count 1 --owner <discord id>
It seems to work. It starts up anyway and does PUTs and GETs.
I had to use my own peer to start it though. Whatever is wrong with the standard peers and the service definition is still an issue.
1 Like
Yes, so far this is for advanced users only. The launchpad should not pick this up yet.
Which peers are you not able to connect to? Just for clarity (ie, initial network-contacts file or the new one)
This is where we think it’s the relay funtcionality that’s hampering them, which this release will aim to solve
2 Likes
It’s the standard network peers. There’s just something wrong with the service definition. It might be that it’s too long. Or something else. A couple of days ago I thought it was that the list of peers isn’t enclosed by " "s but that doesn’t seem to be it.
But it does start if I just specify a single peer which is my own.
2 Likes
There have been two iterations of these, so can you provide the first peer in the contact list you’re seeing please?
Ah, yeh, if it’s adding all of the peers it will be. That’s something fixed in an upcoming release (soon 
)
2 Likes
Apparently that’s someone trying out the latest node binary 
2 Likes
there is a error there. nobody has 15k nodes 
7 Likes
If someone making that chart with their nodes took them down… 
2 Likes
ahhh bingo. also why I eased out, I realised that it was a public monitor of me 
5 Likes
I noticed increased memory usage… if a specific event caused a RAM usage spike that put some of my nodes over the edge, it may have done the same for many others that didn’t have much spare RAM.
Just a guess!
1 Like
Are you referring to the event noted above? If in the current network state your nodes are going over the edge with small hiccups and minimal data flying about you are running too many. Once data starts moving for real you will go into full blown melt down.
I am worried a great many are loaded up to what is “working” in a idle network. Going to be a mess when things start moving.
8 Likes
Yes. That’s just a possibility I’m raising… may not be the best explanation, but something must have caused half of the network’s nodes to die in a coordinated way.
Checking on my nodes that died, the memory explanation probably isn’t correct, as they’re only using around 20-25% of their RAM… I thought they had less than they do. CPU wise they should be fine as well, as they’re storage-bound.
I have noticed memory creeping up on some VPS’ I’ve been running.
What do you think might caused the above node-crash event (assuming it’s not an issue with the device doing the logging and there was no crash of half of the nodes?)?
1 Like
almost certain it was a monitoring event not half a network crash.
6 Likes
RPC Socket is not what is required for outside nodes to connect right? Only for internal communication for apps and such… It doesnt seem to list the nodes port in the details… probably required.
2 Likes
That is in inevitable, everybody wants more Nanos 
I think it won’t be that bad, because current network is big and empty, so impact of more data uploads won’t be that significant as in previous testnets and churn will be fast.
5 Likes
Yeah I started gathering that much hearing everyone talk about how many nodes has been comfortable so far and I am holding off adding any until I know what mine can handle when actually busy.
2 Likes
Yes and no.
RPC is remote procedure call. Its a way to talk to application/services that have no terminal. A monitoring program can issue rpc calls to get status, and/or control the application.
Many servers have rpc calls going over the network.
BUT for us no one is doing that and its for applications on your local machine like launchpad and safenode-monitor
1 Like