fix(scenarios): retry RPC query in compute-target-height (chain-RPC bind race)#295
Merged
Merged
Conversation
…ind race) Live manual run on harbor surfaced this: `seictl nd watch --until=Ready` returns when SeiNode pods report Running (status.phase=Running, all plan tasks Complete) — but seid's Tendermint RPC server takes a few more seconds AFTER that to actually bind port 26657. The compute-target-height bash step's single-shot curl loses that race: curl: (7) Failed to connect to <snd>-internal.nightly.svc:26657 after 10 ms: Could not connect to server failed to parse latest_block_height from .../status Manual curl from a fresh pod 90s later returns HTTP 200 with height 286 — the chain IS up, just not at the instant `nd watch` returned. Wrap the curl in a 30-attempt retry loop with 3s sleep (90s window) and a 3s --connect-timeout. Matches the retry pattern resolve-proposal-id already uses (different step, same shape: chain-side query that needs to tolerate a brief warmup). Symptom chain on the live run: compute-target-height exits 1 → workflow-vars ConfigMap not created → downstream submit-upgrade-proposal seitask-runner pod stuck in CreateContainerConfigError because its envFrom configMapRef can't resolve. Single fix at the source resolves the whole cascade. Bug #9 in the major-upgrade-runs-end-to-end debugging chain. Same shape as several earlier ones: an assumption about timing/readiness that doesn't survive contact with the cluster.
|
You have used all Bugbot PR reviews included in your free trial for your GitHub account on this workspace. To continue using Bugbot reviews, enable Bugbot for your team in the Cursor dashboard. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Live manual run on harbor: `seictl nd watch --until=Ready` returns when SeiNode pods report Running, but seid's Tendermint RPC server takes a few more seconds AFTER that to bind port 26657. The single-shot curl in `compute-target-height` loses that race:
```
curl: (7) Failed to connect to -internal.nightly.svc:26657 after 10 ms:
Could not connect to server
failed to parse latest_block_height from .../status
```
Manual curl from a fresh pod 90s later returns HTTP 200 with `height: 286` — the chain IS up, just not at the instant `nd watch` returned.
Fix
Wrap the curl in a 30-attempt retry loop with 3s sleep (90s window) + 3s `--connect-timeout` per attempt. Matches the retry pattern `resolve-proposal-id` already uses (different step, same shape: chain-side query that tolerates brief warmup).
Symptom chain
Single fix at the source resolves the whole cascade.
Should `seictl nd watch --until=Ready` itself wait for RPC?
Probably. The SeiNode controller's Ready signal currently means "pods are scheduled and plan is complete." A "RPC bound" stricter condition would push this retry logic into the controller and let scenarios stay single-shot. Not in scope here — file as a separate follow-up.
Bug #9 in the major-upgrade-runs-end-to-end debugging chain. Same shape as several earlier ones: a timing/readiness assumption that doesn't hold under load.
🤖 Generated with Claude Code