Skip to content
This repository has been archived by the owner on Nov 30, 2018. It is now read-only.

The aursec-init process is taking a long time #28

Open
netlexer opened this issue Mar 1, 2018 · 13 comments
Open

The aursec-init process is taking a long time #28

netlexer opened this issue Mar 1, 2018 · 13 comments
Assignees

Comments

@netlexer
Copy link

netlexer commented Mar 1, 2018

I'm just curious, how long does the aursec-init process take? It's been over six hours already, on a HP core i5 laptop.

@Lifree
Copy link
Collaborator

Lifree commented Mar 1, 2018

Hi netlexer,
the init process should take less than an half an hour... are you connected to the internet to find other nodes? What was the last output of the script?

@netlexer
Copy link
Author

netlexer commented Mar 1, 2018

I started it up like this:

$ sudo aursec-init
Generate DAG: 100%
Wait for synchronisation

and it's remained like this. I'm definitely connected to the internet.

@netlexer
Copy link
Author

netlexer commented Mar 1, 2018

OK I killed the process, removed /var/aursec and started up aursec-init with the --verbose flag. It's currently printing 'Getting current block' in a loop.

@Lifree
Copy link
Collaborator

Lifree commented Mar 1, 2018

As i thought ... it cant connect to our network... i will try to reproduce your problem

@netlexer
Copy link
Author

netlexer commented Mar 1, 2018

Thanks. Let me know if there's anything I can do here.

@Lifree Lifree self-assigned this Mar 1, 2018
@Lifree
Copy link
Collaborator

Lifree commented Mar 2, 2018

We have found the problem (our bootnode is overloaded) and work on a fix.

@clawoflight
Copy link
Owner

The bootnode being the server that initiates the p2p connections. Without it, users can't find each other.

@clawoflight
Copy link
Owner

@netlexer Peer discovery should work again, can you confirm?

@netlexer
Copy link
Author

netlexer commented Mar 5, 2018

Well it's been running for only about 10 minutes, so this may be premature... it's still waiting for synchronisation and printing 'Getting current block' (about 40 times so far). However, there is an established connection made by geth, which wasn't there before:

$ sudo netstat -tuapw|grep geth                                                                                       [ 8:28am]
tcp        0      0 localhost.localdom:8105 0.0.0.0:*               LISTEN      18761/geth          
tcp        0      0 kore:33776              139.59.130.248:30200    ESTABLISHED 18761/geth          
tcp6       0      0 [::]:30200              [::]:*                  LISTEN      18761/geth          
udp6       0      0 [::]:30200              [::]:*                              18761/geth    

This was from starting aursec-init -v again.

@clawoflight
Copy link
Owner

It does take a while, just not 6 hours :)

@netlexer
Copy link
Author

netlexer commented Mar 5, 2018

It's still running I'm afraid. I've created a paste from the journal output (about 60 lines). It's at https://paste.ee/p/1krMQ - lines 1 and 3 - "couldn't add port mapping" - not sure why my router couldn't add this, but I added the port 30200 forwarding manually... it didn't make a difference.

@netlexer
Copy link
Author

netlexer commented Mar 7, 2018

I'm curious, how many participants are there in this project? Also, is there a way of determining how many package hashes have actually been stored on the blockchain? I'm interested in this, but have found virtually no references to it in the forums or on the wiki. Thanks.

@clawoflight
Copy link
Owner

Have you tried restarting it? We confirmed that it works in vagrant, so it should work for you, as well.

This was @Lifree and my Bachelor's thesis. We are not sure how many users we have because downloads are not a very useful metric, and ethereum wallet addresses change whenever someone re-inits. We know there's more than us, but probably not very many.

We ship aursec-tui, which allows inspecting hash submissions to the blockchain, and obviously you can fetch the submission count for a package id + hash combination.
We don't have aggregated metrics, but a total submission counter would be useful and fairly straightforward to implement, I'll open an issue for that :)

We announced this on aur-general, but didn't get much response. We are not active on the forums - I spend most of my time in the MLs or IRC.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants