Commit Graph
8 Commits
Author SHA1 Message Date
Sheya Bernstein d8b8f0dffd fix(helm): add missing app.kubernetes.io/instance label to volume service (#8403) 2026-02-22 07:20:38 -08:00
Sheya Bernstein 844859de7f fix: add filer fallback after consecutive connection failures (#8000) 2026-01-11 12:14:27 -08:00
Sheya Bernstein 8740a087b9 fix: apply tpl function to all component extraEnvironmentVars (#8001) 2026-01-11 12:14:16 -08:00
Sheya Bernstein c0188db7cc chart: Set admin metrics port to http port (#7936)
* chart: Set admin metrics port to http port

* remove metrics reference
2026-01-02 12:15:33 -08:00
Sheya Bernstein 6f28cb7f87 helm: Support multiple hosts for S3 ingress (#7931) 2026-01-01 07:41:53 -08:00
Sheya BernsteinandChris Lu 915a7d4a54 feat: Add probes to worker service (#7896)
* feat: Add probes to worker service

* feat: Add probes to worker service

* Merge branch 'master' into pr/7896

* refactor

---------

Co-authored-by: Chris Lu <chris.lu@gmail.com>
2025-12-27 13:40:05 -08:00
Sheya Bernstein 7f611f5d3a fix: Correct admin server port in Helm worker deployment (#7872)
The worker deployment was incorrectly passing the admin gRPC port (33646)
to the -admin flag. However, the SeaweedFS worker command automatically
calculates the gRPC port by adding 10000 to the HTTP port provided.

This caused workers to attempt connection to port 43646 (33646 + 10000)
instead of the correct gRPC port 33646 (23646 + 10000).

Changes:
- Update worker-deployment.yaml to use admin.port instead of admin.grpcPort
- Workers now correctly connect to admin HTTP port, allowing the binary
  to calculate the gRPC port automatically

Fixes workers failing with:
"dial tcp <admin-ip>:43646: connect: no route to host"

Related:
- Worker code: weed/pb/grpc_client_server.go:272 (grpcPort = port + 10000)
- Worker docs: weed/command/worker.go:36 (admin HTTP port + 10000)
2025-12-24 12:22:37 -08:00
Sheya Bernstein 911aca74f3 Support volume server ID in Helm chart (#7867)
helm: Support volume server ID
2025-12-24 10:52:40 -08:00