Sanjay

@sanjay

216 Following

2942 Followers

Sanjay pfp

Yes, that willl work. However, the fact that the page_token is the rocksdb key is an implementation detail. You shouldn't rely on it. It's meant to be opaque to the consumer, we can't guarantee it won't change in the future. If you're ok with that risk, it's an acceptable use.

0 reply

0 recast

1 reaction

Sanjay pfp

We can't guarantee the pagination is stable because older messages could be deleted/pruned/revoked. Depending on how frequently you're doing this, you might want to consider using the events subscribe endpoint (you can filter by fid). With snapchain, you can also use block numbers.

1 reply

0 recast

0 reaction

Varun Srinivasan pfp

Varun Srinivasan

Would you be interested in running a beta snapchain node? Please reply if you want to be in the beta

71 replies

14 recasts

96 reactions

Sanjay pfp

It deletes all messages signed by that signer. So follows might be deleted, but does not impact followers.

0 reply

0 recast

4 reactions

Sanjay pfp

Probably the most challenging and fun project I’ve ever worked on. Incredible work by @dynemyte @suurkivi and @cassie. Special shoutout to the Informal Systems team for building Malachite, a rock solid rust Tendermint library which powers snapchain.

8 replies

19 recasts

125 reactions

Sanjay pfp

It's not validated by the hubs, so consumers have to check the app the signer is associated with and decide if they trust that the app has validated the connection. See the FIP: https://github.com/farcasterxyz/protocol/discussions/199.

1 reply

0 recast

0 reaction

Sanjay pfp

Not at the moment, but this is something we can explore with snapchain read nodes

0 reply

0 recast

2 reactions

Sanjay pfp

I haven’t tested it, but no reason it shouldn’t work. It just needs the standard ETH RPC apis.

0 reply

0 recast

1 reaction

Sanjay pfp

Yes, we should be able to for read nodes.

0 reply

0 recast

1 reaction

Sanjay pfp

No, none that I know of. I use an M2 for development and it's worked fine.

0 reply

0 recast

1 reaction

Sanjay pfp

Looks like it can’t connect to the hub? Is it running? Is the hub url correct? I think the port should be 2283?

1 reply

0 recast

1 reaction

Sanjay pfp

Hmm, people are out today. I don't have an ETA for the fix. I suggest working around this for now by disabling snapshot sync. I'll set up on some alerts so we can catch this before the snapshot expires next time.

1 reply

0 recast

0 reaction

Sanjay pfp

1 reply

0 recast

1 reaction

Sanjay pfp

Hmm, this is odd that the messages are increasing so fast but peers are 0. Wonder if it's more of a metrics reporting issue than the fact that there are no peers. Could you DM me the most recent logs from the hub?

1 reply

0 recast

0 reaction

Sanjay pfp

Try now? https://download.farcaster.xyz/snapshots/MAINNET/DB_SCHEMA_11/latest.json points to the latest snapshot

1 reply

0 recast

0 reaction

Sanjay pfp

Not that if it's a brand new hub, I think it will still fail. The flag is only for catchup sync. You have to pass the `--disable-snapshot-sync` cli flag to bypass snapshot entirely (and it will start with no data). I would wait for the snapshots to be fixed if it's a new hub.

2 replies

0 recast

2 reactions

Sanjay pfp

Yes, good point. That will work too.

1 reply

0 recast

0 reaction

Sanjay pfp

Look like there's an issue with the hub that generates the snapshots. We're fixing, hope to have the latest snapshot uploaded later today.

1 reply

0 recast

2 reactions

Sanjay pfp

We found about ~1K users in this state, we'll fix. Thanks for reporting.

1 reply

0 recast

3 reactions

Sanjay pfp

We've been so focused on snapchain, I forgot to do the Hub protocol release for Nov 27. Just released Hubble 1.17. Please upgrade before current version expires at Dec 11 midnight UTC. Apologies for the short notice!

0 reply

12 recasts

93 reactions