How to proper sync Nitro with Prysm beacon chain?

[xbin0r]

Hello, I was runing nitro node for quite a year, but suddenly when new updates came - and I forgot to update in time, my db was corrupted.
I tried repeating same process as always - downloading the snapshot and use it but well latest nitro node is not compatible with my Prysm beacon chain database - im missing blobs or sidecars whatever.

Can somebody help me to fix my Prysm beacon chain node to provide all the data the nitro node needs?
Im desperate and discord is full of scammers.
Thank you very much for any help provided.

[gligneul]

In the last Ethereum main hard fork (Fusaka), a new feature called PeerDAS was introduced. With this feature, L1 consensus nodes don't store by default all the blobs necessary to run L2s, such as Arb1. What you could do is run Prysm with the --semi-supernode flag, which downloads all the data necessary for reconstructing the blobs Arb1 needs. Please take a look at Prysm documentation: https://prysm.offchainlabs.com/docs/learn/concepts/blobs/#post-fusaka

[xbin0r]

Ok so I dont need archival prysm node with slots (blocks) from 0 and super node ? semi-supernode is enough ?

[gligneul]

Archival + semi-supernode if you want to sync Nitro from Genesis. If you run Prysm with only semi-supernode, it will retain blobs from the last 18 days only, which might be enough if you start Nitro from a recent snapshot. You can also set --blob-retention-epochs in Prysm to store the blobs for a period longer than 18 days.

[xbin0r]

Thanks for clarification. Im planning to run nitro archival too - does it need Prysm archival node aswell ?
Im trying to sync the prysm archival for over few weeks already but im stuck on 8.62m slot height

[gligneul]

Yes, syncing a Nitro archival requires all the blobs since genesis, so you need a Prysm archival node. Note that after the Nitro node is up to date, it doesn't need to access old blobs anymore, so a regular Prysm node is enough. (You still need to use the --semi-supernode flag.)

About the Prysm issue, I suggest you ask in the Prysm Discord channel: https://discord.com/invite/prysm

[xbin0r]

@gligneul Thank you for your reply and advice. I joined the prysm discord, it seems runing an archival prysm node is pretty hard. I will need someone to provide me with datadir or something, cause im not able to get over some slot height over days already.
I would love to setup Arbitrum archival too but for this i need full prysm chain.
Maybe you know someone with resources? I might pay for whole prysm chain - no problem.
Thank you again.

[gligneul]

Instead of running your own beacon node (Prysm), you could use an external RPC provider. In our docs, we have a list of a few providers that have historical blob data, so you can run an archival Arbitrum1 node: https://docs.arbitrum.io/run-arbitrum-node/l1-ethereum-beacon-chain-rpc-providers

[xbin0r]

Finally I have whole Prysm datadir from on of the Prysm guys from their discord - thanks for providing the url ;)
Thanks for help and my last question is, how big is full Arbitrum Nitro archival datadir? So I can prepare for it, thanks!

[gligneul]

Currently, a full sync of Nitro archival is ~25tb, maybe a bit more. We are currently working to reduce this by enabling pathdb support, but it is still in progress.