[General] Specific snapshot for external node recovery

[donavernokg]

Team or Project

No response

ZK chain

Era

Environment

Mainnet

L2 block number

No response

Provide a brief description of the functionality you're trying to implement and the issue you are running into.

I am using external node for L2 chain. I would like some historical data from my node, but it was pruned. Now I am trying to figure out how to restore from a particular snapshot so I don't have to start syncing from genesis. I also do not have the space to use a DB dump. What is my best option? Can I select a specific snapshot before the block I want, so I can sync more conveniently?

Repo Link (Optional)

No response

Additional Details

No response

[kaymomin]

Hey @donavernokg, this documentation on pruning and snapshots covers this topic.

• If you only need recent state and not the entire rollup history, you can restore your node from a snapshot taken before the block you’re interested in. After restoring, your node will sync forward from that snapshot, which is much faster and uses less space than syncing from genesis.

• However, if you need access to the entire rollup history, the only option is to use a full Postgres database dump and disable pruning.

[kaymomin]

I've asked the team to confirm you the above. Will get back to you soon.

[kaymomin]

@donavernokg only the statement "after syncing the node, it would have been pruned" is not totally accurate.

The node prunes its history asynchronously (it can take hours or even days to finish pruning existing history after it has finished syncing, depending on your hardware).

[donavernokg]

Thank you, I am attempting to extract the dump to my existing fresh node using the below. Not sure how long it will take, hopefully it works:

• docker compose up -d postgres

• docker compose up -d external-node

• sleep 10

• docker compose stop external-node

• docker compose exec postgres
  psql -U postgres
  -c "DROP DATABASE IF EXISTS zksync_local_ext_node"
  -c "CREATE DATABASE zksync_local_ext_node"

• docker compose exec -T postgres pg_restore
  -U postgres -d zksync_local_ext_node
  < external_node_backup.pgdump

[donavernokg]

@kaymomin hello, what is the total storage needed for restoring a fresh node from PG dump currently? The docs stated 3 TB + 8 TB, what is the actual size needed now? How long should it take based on the doc's system requirements, an estimation? Is there a way for the pg restore to be quicker?

[donavernokg]

I executed this command for my pg_restore

nohup pv external_node_backup.pgdump | docker compose exec -T postgres pg_restore -U postgres -d zksync_local_ext_node -v < external_node_backup.pgdump &

and received this error, after a long while. I thought it was a nohup issue, but I tried exiting completely and reentering my shell and it was still running.

mainnet-node-postgres-1  | 2025-06-28 21:12:21.187 UTC [701] ERROR:  canceling statement due to user request
mainnet-node-postgres-1  | 2025-06-28 21:12:21.187 UTC [701] STATEMENT:  COPY "public"."call_traces" ("tx_hash", "call_trace") FROM stdin;
mainnet-node-postgres-1  | 	
mainnet-node-postgres-1  | 2025-06-28 21:12:21.191 UTC [701] LOG:  could not send data to client: Broken pipe
mainnet-node-postgres-1  | 2025-06-28 21:12:21.191 UTC [701] FATAL:  connection to client lost