Repair improvements

[raymondjacobson]

• Mediorum / update should use peers from eth service instead of its own fetching mechanism
• Remove artificial delay for syncing files
• Sync files reverse chronologically instead of chronologically
• Expose EthService to http
• Try to repair from peers even if we can't find R many

I left it out of this PR, but I think we can probably make repairs happen concurrently. They're very nerfed atm.

Testing:

• verified logs that peering is working and mediorum sees other nodes
• curl https://node1.oap.devnet/eth.v1.EthService/GetRegisteredEndpoints | jq

make up

go run ./examples/upload/main.go  

# copy returned cid

curl http://node1.oap.devnet/internal/blobs/info/baeaaaiqseaxzf3hjvuab2d7xyj3qx7whd4w3mpa2xwcprl4okiqc3whqlxjzc
curl http://node2.oap.devnet/internal/blobs/info/baeaaaiqseaxzf3hjvuab2d7xyj3qx7whd4w3mpa2xwcprl4okiqc3whqlxjzc
curl http://node3.oap.devnet/internal/blobs/info/baeaaaiqseaxzf3hjvuab2d7xyj3qx7whd4w3mpa2xwcprl4okiqc3whqlxjzc
curl http://node4.oap.devnet/internal/blobs/info/baeaaaiqseaxzf3hjvuab2d7xyj3qx7whd4w3mpa2xwcprl4okiqc3whqlxjzc

[rickyrombo]

mind changing this ticker to be once every 24 hours to cut or RPC usage?

[raymondjacobson]

this just pulls from the eth service actually though, so i dont think it touches RPC at all

[raymondjacobson]

just reading from the db

go-openaudio/pkg/eth/db/sql/reads.sql

Line 2 in 125f33f

select * from eth_registered_endpoints;