Autonomi Network Launch Announcement

yapp - did reset “many” nodes lately and I’d state the network is at ~7.2k nodes actually

2 Likes

guess that means we’re somewhere around 2 €/node per month now for cloud nodes (~6 eMaid) … well :smiley: … and raspberry pi’s don’t seem to be suitable for running nodes anymore it seems xD @Bux @JimCollinson

on the upside … with that little nodes per machine nobody will be running out of disk space anymore :smiley:

2 Likes

Mine are all well established so “should” be accurate, I wonder why @peca sees many more.

Any idea how data holds? @aatonnomicc?

1 Like

interesting iv got two identical machines running at home on a miktotic router
both are only running 5 nodes each.

1 was restarted about 5 hours ago is seeing

2 running from the start of the network is seeing


6 Likes

Well that solves it then :rofl:

How are your uploads doing? Gets still good? Sorry I am on phone.

2 Likes

still trying to upload Linux mint iso but haven’t tried a download in a day or two.
was kind of hoping the next release will get things back to normal speed for uploads and downloads :slight_smile:

4 Likes

I said to myself I’ll keep it going for a few months to help with launch, I know I’m going to lose out but hopefully it’s helping.

6 Likes

The CPU usage of the nodes is increasing. I’ll have to reduce the number of nodes…

2 Likes

My 2 nodes are soldiering along still. System load seems to be up at least 50% though.

5 Likes

I am only seeing the 7K to 8K nodes size.
Also about 1 or 2 nodes out of 25 nodes a day are growing beyond 1GB in memory size. They get clobbered and replaced.

5 Likes

Is the plan to update only or are restarts still on the table.
I am down to 75 nodes. 🫨

2 Likes

It will depend on the change.

At this time it is possible to do resets and sometimes it is the quickest way to implement a breaking change rather than some complex set of upgrades

3 Likes

For me, after the initial node die-off, my node count stabilized at 58 from 150 initially…


Check out the Dev Forum

2 Likes

Amongst others I have 3 identical machines (8core Ryzen, 32 GB RAM), in non-ethereum testnets each of them ran 150+ nodes no problem.
With this version I started with 50 per machine. Two of them are running from the start and are down to 38 an 37 nodes (30-40% CPU usage). The third machine had a “meltdown” and dropped to 6 nodes. I tried to reset, but I am not able to get above 10 nodes. It is 95% idle, but as soon as I add another node it goes crazy on CPU for a while and drops nodes.

I haven’t yet resorted to running nodes without node-manager, but the current overprovisioning protection is too aggressive.

4 Likes

(and the cpu spikes caused by the node manager itself don’t exactly help … may help to start the nodes via systemd directly)

3 Likes

In my case, the current network not only increases CPU consumption but also destroys my VPN connection.
I have had to reduce the number of nodes by almost half, compared to other testnet, to make it work.

1 Like

Well I have my flowerpot now.
I have 45 nodes running.

The real cracker here is I started with 65 devices. All 4 core 8gb ram.
So magically some have lost all nodes.
I was previously able to run 10-15 per per device.

4 Likes

I’ve seen one occasion a couple hours ago where I added a node via node manager, the service was created but node manager didn’t list it on status view and couldn’t start it (and other adds fail because the node manager wants to create the exact same service…)

is it possible this is the case an your 0-node runners? service file for safenod1 existing but node-manager telling there’s no nodes that can be started (?)

I have been finding issues with the contacts file not being accessible much of the time. This causes the node to fail after retrying 7 times to retrieve it. That takes between 30 and 60 seconds to finally fail

4 Likes

I am not home and don’t have ssh set up.

So I can only guess. But there has been a gradual decline since the start.

Before I left home I already had some down to 2 nodes and a few returned a rpc error which I pressume had none.

I had no time to investigate.

I have no fancy scripts adding nodes automatically, only one that removes stopped nodes completely.

3 Likes