I think it’s interesting that my new nodes got hammered, but my existing nodes (all have been updated) haven’t seen anything thrown at them.
2 Likes
seeing the same thing here
1 Like
New nodes were filling up but my original ten were not doing anything so I killed those and started again and these show immediate activity, but no relevant records 
:
4 Likes
Had to do the same.
A lot of puts as soon as new nodes were spinned. No nano’s so far, but I guess kind of normal if there’re no new uploads in the network due to lack of coins… (please correct me if I’m wrong).
Let’s see if I manage to earn some fake nanos overnight, have 20 nodes in a rpi4.
2 Likes
yea mass restart on my side many nodes joining over next few hours
1 Like
I’ve been working my way down in value this whole time. Down to 0.00000005 and it’s still not going through. 
2 Likes
Obviously a major malfunction. Everybody stay by their screens.
1 Like
Our nodes have been upgraded to 0.105.6.
5 Likes
I stopped, then started my nodes via safenode-manager and they look like they now have peers! It looks like that fixed my issues. Thanks!
2 Likes
On the down side, I can no longer download the images for my blog or RustRover-233.14015.155.tar.gz again.
safe files download "RustRover-233.14015.155.tar.gz.3" c867ee5142bbe968818c8ca2bdef5f53a97487e4ec8652c3a65a368a0ac1f356 -b 1
Logging to directory: "/home/paul/.local/share/safe/client/logs/log_2024-04-10_19-44-41"
Built with git version: f532de4 / stable / f532de4 / 2024-04-10
Instantiating a SAFE client...
Trying to fetch the bootstrap peers from https://sn-testnet.s3.eu-west-2.amazonaws.com/network-contacts
Connecting to the network with 50 peers
🔗 Connected to the Network Downloading "RustRover-233.14015.155.tar.gz.3" from c867ee5142bbe968818c8ca2bdef5f53a97487e4ec8652c3a65a368a0ac1f356 with batch-size 1
Error downloading "RustRover-233.14015.155.tar.gz.3": Network Error GetRecord Query Error RecordNotFound.
I have restarted my nodes for a second time. After a couple of hours the second batch were showing PUT/GET activity but no records at all. Immediately after restarting (same safenode version) my third batch has one node with 8 records:
4 Likes
I’m trying safenode-manager for the first time. It’s great!
I can see one tweak that is needed: I think we need an option to stagger the startup of nodes. It will be kinder to network connections and low powered hardware. In the past I’ve been starting up nodes with a delay of 7 mins to let the influx of records happen and not swamp the network and the CPU which can cause nodes to go unresponsive.
3 Likes
Need example code on how to remove nodes from safenode-manager.
‘safenode-manager (stop/remove) --service-name=(safenodeXX)’ xx is the number you want to remove
1 Like
Indeed. I thought this was a feature that was getting added RealSoonNow.
Its growing on me but I’m still uncomfortable using root/sudo for what seems routine jobs. I know elevated privs are required to start/stop services.
1 Like
We have it for starting nodes --interval 5000
Equals 5 seconds
This is only for initial start though right? That will start the initital service. However, on reboot, they all just get started up by init as services. (this is my understanding. I need more time for testing)
I really don’t think this is an issue for “most users” but those of us that are going to be running boxes for the express purpose of supporting the network it will probably be an issue.
My planned workaround is allow safenode-manager to create all the service (so it will play nice with the node registry), but then move the services out of normal startup process. Add a service to the normal startup process that will ramp them up much like @Shu has described.
How that happens? Not sure yet. But google knows, so I will soon too. 
1 Like
@chriso - I haven’t played with upgrade action on safenode-manager. Using the safenode-manager from AlphaToleranceTestNet shows its version as 0.7.4-alpha.1.
I noticed it doesn’t have a --path command here for the upgrade binary in-case you want to use a custom build binary as input, which would prevent say for myself from using it at this time.
I may have to just place the proper safenode binary in each of the safenode’s local snapshot folder where the actual safenode lives via external script for now (between the stop and start operation).
I would also like to know for this testnet/upgrade, what is the branch or tag label that is being used here from Github which I can build the safenode binary from source, so I don’t accidentally pick the wrong one.
I see alpha, stable, and main branches for now. Hmm.
Yes, on prior testnets, I moved the services out of autostart mode for the reset + add operation for now, along with the ramp up based on cpu % for now. Makes life easier.
Note: It will break safenode-manager refresh off the registry (I am forgetting the exact scenario), so you would have to restore the default run levels if you were doing one of the supported operations again with safenode-manager (forgot which one now), but I had to implement a temporary restore of run levels when resetting, then eventually put them back to non automatic after using safenode-manager in the middle of the sequence… So user be aware!
Will have to wait for that rust crate to have the features to allow Chris to default the automatic start mode possibly to off unless a flag is set for automatic start mode.
1 Like
Ah! Didn’t realise that. I’ll try it. But what @wes said about needing it for restarts might still be the case.
2 Likes
You can use the stable branch for this.
If the upgrade doesn’t have a --path, I’ll add that soon.
4 Likes