ClientImprovementNet [22/09/23 Testnet] [Offline]

Hey @joshuef, how can I get failed to fetch... in a failed upload, after uploading the same file (several times) successfully before? Aren’t the new chunks the same as the old ones, and trying to fetch a new chunk should deliver one of the old ones? What am I not getting here?

You’re not not getting anything. In theory.

In practice it could be payments (where the error message needs improved). If you’re failing to pay, that data wont be retreivable, and that may be your issue. If eg your wallet is corrupt (past spend went to the network, but wasn’t written locally), that could be it (the network sees a double spend, so doesn’t accept the new one so you can’t pay for your data again; atm, every spend is done again regardless of if you’ve paid for chunks before).

1 Like

Stupid question from a non-tech guy who tried but failed to connect (i guess because i am behind 2 home routers):

Do files get lost at all in these testnets?

4 Likes

There are no stupid questions here, only the unasked questions are stupid :slight_smile:

Files don’t get lost, the only thing that happens is that they can’t be uploaded to the network for some reason, or if they manage to be uploaded, they are saved in pieces, distributed all over the network.

4 Likes

On the other hand, they do get lost when the test ends and the testnet is taken down.

4 Likes

Yep.

======= Verification complete: all chunks paid and stored =============
Uploaded all chunks in 53 minutes 29 seconds

This is the file if anyone wants to try to download it. Size: 393,170,097 bytes

test01.mkv → 1be692a7bf9096735159b645e92a6c9f0ad6bf42b14d58e647648a4165fecb21

3 Likes

just tried to download no dice

Error downloading "test01.mkv": Chunks error Not all chunks were retrieved, expected 750, retrieved 748, missing [e8f7b6(11101000).., f84701(11111000)..]..

real    10m16.973s
user    2m25.694s
sys     0m42.674s

So, who fancies trying the latest client?

safeup client --version 0.82.3

To try out it’s neatest feature, you’ll need to unset SAFE_PEERS (on ubuntu at least… windows I’m not sure). You should then be able to run the client as usual, but without having to specify peers anywhere.

ie, no SAFE_PEERS=<peerid> and no --peer=<peerid>.

If you do it’ll overwrite the network-contacts retrieval from s3 and it’ll just use those.

8 Likes

Client 82.3 passes the begblag test for me :slight_smile:

ubuntu@safe-hamilton:~$ safe -V
sn_cli 0.82.3
ubuntu@safe-hamilton:~$ safe files download BegBlag.mps 80744b3d25bab269cab54e8baccf4f54f1aa01615230b99171bc3576c1ca7230
Built with git version: 7f609f8 / main / 7f609f8
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 w/peers: ["/ip4/142.93.232.219/tcp/38095/p2p/12D3KooWLAX6Z1m5gNxPGZQRV6VEzCtipNGcP1YhrAYNYT3yq5mv", "/ip4/64.227.158.176/tcp/34893/p2p/12D3KooWG3cHz8aM9Zf2Gyar7NBb2BZ2wfcBf1zY7PHuUtn3EkvL", "/ip4/147.182.237.224/tcp/46429/p2p/12D3KooWFCbRRe6GoyTvgbVqDK7VnD3TwEsM2876pBzuB1G3KHKj", "/ip4/139.59.125.187/tcp/33641/p2p/12D3KooWNUNgbhC2zEUNVJzn72BxD4ooZ7kTE4NSa3pCmcGobGQj", "/ip4/142.93.76.173/tcp/41717/p2p/12D3KooWJ51eYsdyksy87PazS3M7wRVdPNJWfH58u8fxBk2Y9hdp", "/ip4/142.93.76.173/tcp/46229/p2p/12D3KooWQi1eeRbwtRWRvYTSQ2gSKt6M8QjseS65Hwzb7PDGKghE"]...
🔗 Connected to the Network                                                                                                                                                                                                                                         Downloading BegBlag.mps from 80744b3d25bab269cab54e8baccf4f54f1aa01615230b99171bc3576c1ca7230
Client download progress 1/31
Client download progress 2/31
Client download progress 3/31
Client download progress 4/31
Client download progress 5/31
Client download progress 6/31
Client download progress 7/31
Client download progress 8/31
Client download progress 9/31
Client download progress 10/31
Client download progress 11/31
Client download progress 12/31
Client download progress 13/31
Client download progress 14/31
Client download progress 15/31
Client download progress 16/31
Client download progress 17/31
Client download progress 18/31
Client download progress 19/31
Client download progress 20/31
Client download progress 21/31
Client download progress 22/31
Client download progress 23/31
Client download progress 24/31
Client download progress 25/31
Client download progress 26/31
Client download progress 27/31
Client download progress 28/31
Client download progress 29/31
Client download progress 30/31
Client download progress 31/31
Client downloaded file in 32.877776755s
Saved BegBlag.mps at /home/ubuntu/.local/share/safe/client/BegBlag.mps
4 Likes

So I gave it a go and it failed:

Error downloading “test.mkv”: Chunks error Not all chunks were retrieved, expected 750, retrieved 746, missing

[e8f7b6(11101000)…, 2138ff(00100001)…, f84701(11111000)…, c8632c(11001000)…]

Interestingly though I’m missing four chunks whereas you were only missing two.

So it does look like a network problem. Any thoughts @joshuef ? Do we need more copies or is data getting deleted?

Maybe others with fast connections can give this go and see if it works for them. More confirmations here could be helpful.

1 Like

so hopefully not stating the obvious here

so far for this network iv ran nodes of 3 different versions and now on a new client version

does this mean that upgrade ability for a running network has been solved ?

4 Likes

BTW, has anyone tried some other functionalities like:

  • Sending money?
  • Creating an updateable file? (Register, or whatever that was…)
2 Likes

Not necessarily.

We’ll need a protocol level update to be sure. But it’s certainly working as intended so far (and is improved to prev networks :+1: )

2 Likes

Neither my laptop or cloud instance have a logs directory under ~/.local/share/safe/client/. Is there something I needed to do to enable them? Ah, seems I have to use --log-output-dest

I just retried with the ‘failing’ SAFE_PEERS and it was taking ages to connect but I hadn’t set --log-output-dest so I retried with that and got logs but unf it now connects immediately :man_facepalming:

1 Like
expected 750, retrieved 748

Just to add, same as @digipl

2 Likes

Rebooted and upgraded to the latest safe - seems to be working fine now.

Connection to the network much faster now too with v0.82.3

4 Likes

In my opinion both node and client could log to a file as a default. That data is valuable and is too easy to miss now.

3 Likes

Yep, updated to 0.82.3, unset SAFE_PEERS, and it bootstrapped from AWS no problem.

Ubuntu 20.04.6 LTS.

6 Likes

Updated the OP with v0.82.3, removed the export peers instruction for Linux/Mac client.

4 Likes

That’s confusing actually … what were the chunk id’s?

1 Like