Provider crash: "client is not running" / balance-checker "context canceled" – waiting on v0.10.6

[VirgilBB]

Support

Description

Provider crashes shortly after startup with:

• received shutdown request err="context canceled" (balance-checker)
• Error: client is not running. Use .Start() method to start

Environment

• Provider address: akash1nj6ygd4ggz589ldtt4e7yklazxm9zpp8cf7yh7
• Provider URI: https://provider.cerebro.host:8443
• Version: provider-services v0.10.5 (image: ghcr.io/akash-network/provider:0.10.5)
• Mainnet 15, helm chart akash/provider 14.0.6

To Reproduce

1. Provider starts (init completes, cert valid, provider updated on-chain).
2. Provider fetches attributes, starts gRPC on 8444.
3. Balance-checker receives "context canceled"; provider shuts down with "client is not running".
4. Pod enters CrashLoopBackOff.

Error log (relevant lines)

INF starting provider service
INF fetched provider attributes ... provider=akash1nj6ygd4ggz589ldtt4e7yklazxm9zpp8cf7yh7
INF grpc listening on "0.0.0.0:8444"
DBG received shutdown request err="context canceled" cmp=balance-checker module=provider-service
INF shutting down services
Error: client is not running. Use .Start() method to start

Expected behavior

Provider should stay running; balance-checker should not trigger premature shutdown.

Status (from Akash team)

• Fix is in provider#359 (error handling for c.bus.Subscribe()).
• There is no released image with that fix yet.
• v0.10.6 is on RC and will include this fix; waiting for full release.

Request

• Use this issue to track the problem until v0.10.6 is released.
• When v0.10.6 (or the recommended image tag) is available, we will upgrade and confirm the provider stays up.

Related

• Original crash (certs.go:365) was reported and cleared by the dev team; this is the follow-on balance-checker crash.
• Previous console sync issue: #1148 (resolved).

[VirgilBB]

I got a notification that a triage/label workflow failed on this issue (label-actions / label-triage config).