1 Like
This makes me edgy though.
How do you guys calculate what you are able to run?
8GB RAM (in my case)
Set 1GB aside for system.
70 nodes then gets 7GB and issues will arise at around 100MB of use per node.
I know that is very rough.
At the end of the day all that matters to me is not losing nodes (long) before Maidsafe does.
1 Like
Iv been running 50 nodes per system on 4 core 6gb ram with 80Gb SSD
3 Likes
Josh did mention another find/fix so an upper limit of 100MB per node may be way high in the next test.
2 Likes
Better safe than sorry can always start more nodes later
1 Like
Perfect, thanks @mav
1 Like
Does the UK have data caps… or no caps “with fair usage policy”
From what I can find luckily my ISP does not have any stipulations but I may soon make them reconsider 
1 Like
Garbage of course.
It is easy to make random file which can be easily validated.
But it is boring.
3 Likes
@stout77 be aware that the safe CLI command won’t find your node wallets if you don’t use the default location. 
That’s not a problem unless you want to check earnings in the wallet for a node.
2 Likes
We’re waiting on a patch release from libp2p for that other fix atm.
9 Likes
I grabbed myself some Charlie Chaplin, there lots of other stuff too.
5 Likes
Ia this patch significant enough to be worth delaying the next testnet for?
Is there any ETA on this?
Should we all go camp on the libp2p Github pages? 
1 Like
Heresy!! Burn him at the stake.
1 Like
Guy fox night is coming up soon looks like we have found our Guy
Thank Josh.
Do you guys have or forsee issues with transfer limits on DO?
2VCPU4GB is 4TB so 20 nodes 200GB, are you tracking it at all? I have not been paying attention.
no
no, but i believe “soon”
definitely not
I have not been! But it’s something to keep an eye on for sure! I\ll add it to my routine (or try to!)
6 Likes
IF I was feeling like a bit of self harm and made it so that others can add their nodes to my online version of ntracking would anyone do so?
It is a headache with migraine potential so def not planned but an idea I keep revisiting.
6 Likes
I pity your migraine already! And your AWS bill…
I think it could be really good though.
I was also thinking about a community syslog server. Imagine if all logs from people participating were already available for Safe staffers and us without recourse to attaching logs to the forum or even uploading as files into Safe itself. But the bill for that could be jaw dropping!
I was also fantasising about a community ELK stack. Imagine the insights we could get from that! But the bill for that would be eye watering!
7 Likes
Someone made a post in either this or another recent test about home routers ability to handle significant tcp connections. Cant find it, am I the only one that finds search on here sub par, anyway not the point.
Another post I can’t find is one by David about the amount of nodes a large machine can potentially run.
For context 300 nodes can yield ~120K connections excluding spikes that do occur.
Edit: the above number is incorrect.
The mean is 280 connections per node so 300 nodes would average 84,000 connections.
I will be tracking latency and bandwidth in the next test and might push my router to see where it starts to buckle.
The way I see it RAM, internet connection and router will be the 3 key deciding factors on how many nodes a user can run.
It seems abundant disk space is far from the top of the list in importance.
Interested in others thoughts, findings so far?
5 Likes
It was me. Yes, I think number of connections will be reached for a lot of people’s home setups way before bandwidth usage becomes an issue. I’ll find the post when I’m back because there was some interesting discussion with at least one other person.
3 Likes