Intel is outperforming AMD on this one.
My Ryzen9 was idling with more nodes last run when my Intels were working way harder for less nodes.
Now it is running less nodes than the Intels.
Intel is outperforming AMD on this one.
My Ryzen9 was idling with more nodes last run when my Intels were working way harder for less nodes.
Now it is running less nodes than the Intels.
My intel is running about half as many nodes
I will test my AMD 24 core machine later when I get around to starting it up. Might make 4 eth addresses first for its nodes. It’ll run the same number since the limits were bandwidth. It sat at <10% CPU with 150 nodes previously
Updated. Thank you
Not sure if you guys should mess with it.
There were 2 fairly decent network drops last night and for the first time ever none of my nodes were affected by churn.
Yes cpu is higher but I’ll take that if it is stable any day of the week.
I saw zero effect on my nodes.
The hot fix is deemed as a critical update by everyone on the team and Chris is actively working on releasing that update based on an internal analysis off the current metrics on the current production binary version out in the field. There is no immediate ETA on when it will be released, but the work is in progress to get it out to the public.
Its meant to reduce CPU usage, memory, active # of open connections, and # connected peers, along with other benefits on multiple other metrics we are carefully observing.
Where can you see / find this? https://network-size.autonomi.space/ doesn’t seem to be working yet.
I’m having various trouble uploading from the CLI.
I did manage to upload a test OGG file, but it returned a non-XOR address from the CLI:
Upload completed in 90.103176159s
Successfully uploaded: ./Free_Test_Data_10MB_OGG.ogg
At address: 1017866621646572147
Number of chunks uploaded: 6
Total cost: 18 AttoTokens
It also reported a bunch of DialErrors prior to success.
I’m guessing that 1017866621646572147 is a raw output, rather than an XOR address, but not looked into that. Either way, it makes it hard to know what the address is to confirm the upload.
When trying to upload the current directory (ant file upload ./
), I always get this error though:
The application panicked (crashed).
Message: Failed to get file/dir name
Location: autonomi/src/client/files/mod.rs:25
Uploading to a named subdirectory seems to get further, but getting varous network related errors:
Error:
0: Failed to upload dir and archive
1: Failed to upload file
2: Error occurred during payment.
3: Cost error: MarketPriceError(ContractError(TransportError(Transport(Custom(reqwest::Error { kind: Request, url: "https://arb1.arbitrum.io/rpc", source: hyper_util::client::legacy::Error(Connect, ConnectError("dns error", Custom { kind: Uncategorized, error: "failed to lookup address information: Try again" })) })))))
4: Market price error: ContractError(TransportError(Transport(Custom(reqwest::Error { kind: Request, url: "https://arb1.arbitrum.io/rpc", source: hyper_util::client::legacy::Error(Connect, ConnectError("dns error", Custom { kind: Uncategorized, error: "failed to lookup address information: Try again" })) }))))
5: error sending request for url (https://arb1.arbitrum.io/rpc)
6: error sending request for url (https://arb1.arbitrum.io/rpc)
7: client error (Connect)
8: dns error: failed to lookup address information: Try again
9: failed to lookup address information: Try again
and:
Error:
0: Failed to upload dir and archive
1: Failed to upload file
2: Error occurred during payment.
3: Cost error: NotEnoughNodeQuotes(bd4205(10111101).., 2, 5)
4: Not enough node quotes for bd4205(10111101).., got: 2 and need at least 5
I suspect enough retries may get those over the line (I’ll keep retrying).
Note that it seems to hit my network hard when doing the above too. Maybe the two things are related and the client is actually working against itself (causing issues with my router, which causes further issues with the client, etc).
Thanks for the report here.
The hotfix will remove the ugly ‘dial error’ output, but I’m not sure about the other issues. I don’t think we see those on our own uploaders.
I’ve put your post in our Slack channel.
I’m not sure if this is also related to my issue (see here). But could there be a reason why my Hetzner nodes do not experience the same issue?
I tried with -p -x and it gave me an XOR address. Maybe I just always did public uploads? Anyway, it worked and was faster too:
Upload completed in 26.057707433s
Successfully uploaded: imim_with_code/Free_Test_Data_10MB_OGG.ogg
At address: d2c6c9ab08720fb64fbc8caf58ad0c64a3137b5689238fd4a45c4b7bdaf144f2
Number of chunks uploaded: 8
Total cost: 15 AttoTokens
I just know the folk who reset systems.
The hotfix is available now:
Does it download automatically or should I do something special when starting my nodes?
If you have existing nodes, you can upgrade them as usual with antctl upgrade
. However you normally do that.
If you are running new nodes, they will automatically go for the new version.
antctl start
fails with a permissions error for me. [Ubuntu 24.04 LTS, Linode VPS]
I used to see the same with safenode-manager
Running as a normal user, not root.
Anyone else finding this?
Attempting to start antnode1...
Attempting to start antnode2...
Attempting to start antnode3...
Failed to start 3 service(s):
✕ antnode1: Failed to connect to bus: Permission denied
✕ antnode2: Failed to connect to bus: Permission denied
✕ antnode3: Failed to connect to bus: Permission denied
Edit: solved courtesy of Dave on superuser.com:
Check linger status
loginctl show-user yourusername | grep Linger
, enable itloginctl enable-linger yourusername
Good old Dave.
Works fine on MSWIN11 2024, everyone just download, extract, pin nodelaunchpad.exe app to (in my case) MSWIN11 task bar, click and go through the routine, stop, reset, ctrl-G set node count, start.
Thanks @chriso !
It is still killing my router, but hopefully it is working well for others.
Well, either that or the CLI uploads. I’ll give in some time over the next days.
Sorry, I’ve never used the LP, so I can’t really provide support there, but, that looks outdated. The crate name is not sn_node
any more. It’s been changed to ant-node
.