Yeh, vdash may not free up memory. I’d have to look but suspect that’s the case. 
7 Likes
Nope I think you are off the hook here, Mark
this is as before but without vdash and a 50 node baby-fleming just idling.
The memory usage is still going up
4 Likes
So it did settle? What ram size is 88% for you there?
6 Likes
Yes it did settle 88% is around 28.5GB.
5 Likes
Can anyone else confirm this alleged memory leak?
steps to replicate on ubuntu
-
Git pull the latest from GitHub - maidsafe/safe_network: The Safe Network Core. API message definitions, routing and nodes, client core api.
-
cargo build --release -
Copy target/release/safe to ~/.safe/cli
-
Copy target/release/sn_node to ~/.safe/node
-
safe node run-baby-fleming --nodes=50 -
Fire up System Monitor - I like to run it full screen but I have a multi-monitor setup - select Resources tab and click the arrow beside Network to hide that panel and increase graph resolution
-
Wait - no need to actually do anything with the network - and watch your memory slowly get used up.
I am also a bit worried about the CPU usage - seems kinda high for an “idling” system
9 Likes
Ill give it a go in a few hours, if not… tomorrow.
10 Likes
500 MB per node, wasn’t that normal value in testnets?
4 Likes
I looked again and this time 50 nodes has plateaued at 14.7Gb ~45% on an idle network
More data points are needed…
As no sane person is ever likely to run 50 nodes on one machine when we are live, I’m wondering if I am worrying over nothing. I have proved to myself that the network splits as expected so what exactly is the point of running that many nodes again?
Possibly I should ask you all to run say 10 nodes and look carefully at memory usage and extrapolate from there…
The real answer is of course to carefully note resource usage of a single node on a testnet but we cant quite ask that real question just yet.
However CPU usage is still a bit high…
I will run it all again with 25 nodes and see if the CPU usage goes down at all.
10 Likes
Anyone trying for Glasto tickets on Sunday?
Casual SAFE get together 2023? 
4 Likes
Maybe the year after launch 
so we should all book them for 2024 
4 Likes
Been a good few years Idk if my memory is screwed but… what inflation?
3 Likes
I see what you mean now, it creeps steadily.
50 nodes on a laptop with 16GB RAM, was at 35% right after all nodes started and then keeps going at roughly 1% per minute until I killed it.
Guessing it will be barely noticeable with a single node.
8 Likes
Exactly, I should find other things to poke at. 15 nodes behaves fine.
It’s something that probably doesnt need sorted right now but I should maybe file an issue so it isnt forgotten about entirely
7 Likes
They normally start off at around 40 MB in my experience.
2 Likes
Did it start again?
2 Likes
no
I hit reply too soon - that was a draft msg, I need to look at the other possible culprit logs first.
2 Likes
Thx 4 the update Maidsafe devs
Really admire how things are always simplified and done elegantly 
for the quick feedback @Southside @Josh
Keep up the good work super ants 
13 Likes
Cool. Been waiting for this one
6 Likes
Thank you for the heavy work team MaidSafe! I add the translations in the first post 
Privacy. Security. Freedom
11 Likes
Like everyone else, I have a lot of questions and doubts, but I love reading about reducing the complexity of the code, improving the network and speeding up the work of nodes - it makes, however it will be an amazing invention!
Maidsafe team - respect!
8 Likes