aptos archive node setup

[riprsa]

hi, i downloaded archive from genesis on my server, now i have a really big archive. how do i do from now? should i unarchive it before running the node? i am a bit lost

there should be fullnode.yaml with

base:
  role: "full_node"
  data_dir: "/opt/aptos/data"
...

so, /opt/aptos/data is unarchived data from s3 storage, or just the archive? i am afraid to try it, because do not want to corrupt the ledger data

[riprsa]

upd: the archive got corrupted, it ate more than 1.5TB of RAM and got OOM killed. so now i am trying to setup an Indexer with full-node, but I need Indexer to have whole blockchain data. Is that even possible?

[0x-j]

Hey does the doc here helps?
https://aptos.dev/en/network/nodes/full-node

[riprsa]

hello. we want to host an indexer (with all ledger data) for a client, they need it for the least possible latency and custom limits for payload size, timeouts, etc (through our proxies)

[riprsa]

so, is this possible at all? self-hosted indexer with archive data?

[riprsa]

@0xaptosj hi, sorry to bother you. is it possible?

[MrSevyz]

I tried too to start a node from genesis restoring from s3 but it keeps going oom.
In my opinion, you should get your hands on a copy of the db of an indexer and then set up your indexer with this: https://aptos.dev/en/build/indexer/txn-stream/self-hosted#configuration . Unlike the legacy indexer this can be bootstrapped using fast-sync .
PS. If you find a backup of a psql with historical data send me a link .

[0x-j]

https://aptos.dev/en/build/indexer/txn-stream/self-hosted#advanced-usage

here includes the link to download the latest indexer db data if you want to do fast sync, does that work?