fix: beacon depends on validator and bridge #28
Conversation
rootulp
changed the title
docker-compose.yml
Outdated
| depends_on: | ||
| - celestia-network-bridge | ||
| - reth | ||
| celestia-network-validator: | ||
| condition: service_healthy | ||
| celestia-network-bridge: | ||
| condition: service_started | ||
| reth: | ||
| condition: service_started |
There was a problem hiding this comment.
I don't think this makes any logical difference as the celestia-network-bridge doesn't start until the validator is "healthy" anyway.
If you wanted to add a health check to the bridge node, you could try using some lightweight method like /network_head.
There was a problem hiding this comment.
Good point. I tried to add a health check using node.Info or p2p.Info but both required an admin token. Looks like /network_head only needs a read auth token so seems a tad easier.
Ref: https://node-rpc-docs.celestia.org/?version=v0.20.4
Created celestiaorg/celestia-node#4007 so celestia-node exposes an un authenticated status endpoint. In the meantime DevOps shared a workaround that they use:
exec [sh -c AUTH_TOKEN=$(celestia bridge auth read --node.store=/home/celestia 2>/dev/null); curl --max-time 0.5 -s -X POST -H "Content-type: application/json" -H "Accept: application/json" -H "Authorization: Bearer $AUTH_TOKEN" -d '{"id":1,"jsonrpc":"2.0","method":"header.SyncWait","params":[]}' "http://localhost:26658"; if [ "$?" -eq 0 ]; then exit 0; else echo "Catching up"; exit 1; fi ]so I'll try something like that
|
A couple of findings:
|
|
Update the celestia bridge node is starting prematurely. If I retry starting it after the validator has been up for some time, it works. However, on initial launch it doesn't work and fails to start the RPC server. The healthcheck on the validator doesn't seem to be granular enough. |
docker-compose.yml
Outdated
| test: ["CMD-SHELL", "/bin/sh -c 'BLOCK_HEIGHT=$(curl -sf http://localhost:26657/status | jq -r .result.sync_info.latest_block_height); if [ \"$$BLOCK_HEIGHT\" -gt 1 ]; then exit 0; else echo \"Block height too low: $$BLOCK_HEIGHT\"; exit 1; fi'"] | ||
| interval: 1s | ||
| timeout: 5s | ||
| retries: 5 | ||
| retries: 20 |
There was a problem hiding this comment.
[note for reviewers] this makes Docker report this service healthy only if the block height is > 1
|
You can fetch with this thing without setting my fork as upstream. Update: also pushed this branch to upstream
|
thanks! didn't know you could do this |
Closes #24
The whole stack now reliably starts