OFFLINE Will it be a Quicky? (run 4)

Also you can change the hard drive from 50 gb up to 200gb and still be on the free tier

1 Like

So what happened to them?

1 Like

It is wrong direction.

1 Like

I think we surely need the money component to bring up more resources, when network is starting to reach it’s capacity.

Nevertheless, the network should also be able to say ā€œI am fullā€ politely and continue all the other functions except storing more data.

7 Likes

Short answer, idk.
Hoping someone like you that is better at log analysis can pitch in there.

Similar loss occurred in all the runs, each time the remaining nodes were nearly maxed out on memory, hence each successive run had more added as it seemed to be the reason for the loss.
This time however out of the 8GB allocated to each node I never saw any go above a max of around 1700MB, they were actually way below that at the time of things going south.

Obviously throwing more and more resources at it is not the cure.
(Although I kinda feel like adding 1 more run with additional CPU’s just to round off the series, but I am not sure there is much to gain from it)

If anything I should have noted the ips of lost nodes which could be helpful, I did not.
(Learning process for me here)
So perhaps that may be reason for a final run in this series.

Edit: I am wondering if adding nohup to the join command will be of any benefit, idk how many people are closing terminals and abruptly killing nodes unintentionally.

10 Likes

It will be needed after network reach stability without money.
Adding complexity will make bug hunting even harder.

I think that hunting such bugs should be first priority for developers.
But they have their own opinion on this topic.
Will see how it goes.

2 Likes

Using the CLI $ safe node join command should do that for you, so perhaps we should encourage people to use that command instead of invoking the node bin directly (it needs the sn_node binary installed at ~/.safe/node/sn_node).

3 Likes

I don’t believe it is working? is it?

Has anyone successfully joined using safe node join?

I tried it but no dice

2 Likes

Safe node join is not working I tried
Also would it be possible to manually specify the port? So I can set up one port forward for safe

1 Like

I’ve just tried locally and it seems to work fine: $ safe node join --skip-auto-port-forwarding

5 Likes

All CLI commands have a --help arg which provides help for all args supported by each of them. For the join sub-command I guess you are looking for --public-addr arg…

…there is also --network-name which could be used, based on the OP steps: $ safe node join --skip-auto-port-forwarding --network-name comnet

7 Likes

:sweat_smile: hidden in plain sight, thanks @bochaco

6 Likes

It’s great that these little specifications and arguments are being cleared up for everybody to use. Please continue adding the exact commands to the OP for simple copying and pasting when running future test nets. Maybe even a single bash script doing everything, including stuff like rm -rf $HOME/.safe would be good.

5 Likes

Agreed, if anyone has any input or suggestions of any kind prior to the last of these until we see any significant changes please shout.

Probably a little later today.

6 Likes

so next time round

safe node join --log-dir ~/.safe/node/logs --skip-auto-port-forwarding --public-addr <public ip address>:<port>
2 Likes

working on a script now

Wee problem - is https://sn-comnet.s3.eu-west-2.amazonaws.com/node_connection_info.config still valid?

Switching to 'comnet' network...
Creating '/home/willie/.safe/node' folder for network connection info
Fetching 'comnet' network connection information from 'https://sn-comnet.s3.eu-west-2.amazonaws.com/node_connection_info.config' ...
Error: 
   0: expected value at line 1 column 1

EDIT2
Space is huuuuge, really huge but even wee spaces can cause embarassment…

2 Likes
#!/bin/bash
rm -rf $HOME/.safe
curl -so- https://install-safe.s3.eu-west-2.amazonaws.com/install.sh | bash
safe networks add comnet https://sn-comnet.s3.eu-west-2.amazonaws.com/node_connection_info.config
safe networks switch comnet
cd ~/.safe/node/
wget https://github.com/maidsafe/safe_network/releases/download/safe_network-v0.51.7/sn_node-0.51.7-x86_64-unknown-linux-musl.tar.gz
tar xf *.tar.gz
chmod +x sn_node

so far i have this

4 Likes

#!/bin/bash

echo ā€œThis script is only for ā€˜vanilla’ linux systems on standard x86-64 PC hardwareā€

clean up from last testnet

rm -rf /home/$USER/.safe/

grab latest sn_node and prepare it for use

mkdir /tmp/comnet

cd /tmp/comnet

wget https://github.com/maidsafe/safe_network/releases/download/safe_network-v0.51.7/sn_node-0.51.7-x86_64-unknown-linux-musl.tar.gz

tar xvzf sn_node-0.51.7-x86_64-unknown-linux-musl.tar.gz

mv sn_node sn_node-0.51.7

chmod +x sn_node-0.51.7

chown $USER:$USER sn_node-0.51.7

#get sn_cli

cd

curl -so- https://install-safe.s3.eu-west-2.amazonaws.com/install.sh | bash

cp /tmp/comnet/sn_node-0.51.7 /home/$USER/.safe/node/

cd /home/$USER/.safe/node/

#Use a symlink so it is easy to keep various versions of sn_node around to easily switch if req’d

ln -s sn_node-0.51.7 sn_node

#set up the network

safe networks add comnet https://sn-comnet.s3.eu-west-2.amazonaws.com/node_connection_info.config

safe networks switch comnet

#Add a node

RUST_LOG=safe_network=info,qp2p=info ~/.safe/node/sn_node --skip-auto-port-forwarding --network-name comnet --log-dir ~/.safe/node/logs

not really tested yet…

2 Likes

Maybe the script could be named something like josh_runX.sh, where X stands for the run number. Just to make certain things are organized and valid.

1 Like