Skip to main content

How Can I Fix The Delegate Installation Error with Readiness Probe Failure in GKE?

Module

  • Harness Platform

Environment

  • Infrastructure: Kubernetes
  • OS: Linux

Issue

After configuration of the proxy alongside testing the connection using the command

wget -p https://app.harness.io/ -O /dev/null

The response is successfull yet Delegate is not running and fails with the below error message

Readiness probe failed: Unhealthy   Aug 19, 2021, 1:14:57 PM    Aug 19, 2021, 1:34:37 PM    111 
Pulling image harness/delegate:latest Pulling Aug 19, 2021, 1:14:32 PM Aug 19, 2021, 1:14:32 PM 1
Successfully pulled image harness/delegate:latest in 162.963829ms Pulled Aug 19, 2021, 1:14:32 PM Aug 19, 2021, 1:14:32 PM 1
Error

2021-08-19 13:38:33.059 IST

#=#=# ##O#- # ##O=# # #=#=-# # -#O#- # # -=#=# # # -=O#- # # # -=O=# # # # -=O=-# # # # -=O=- # # # # -=O=- # # # # -=O=- # # # # -=O=- # # # # -=O=- # # # # -=O=- # # # # -=O=- # # # # -=O=- # # # # -=O=- # # # # -=O=- # # # # -=O=- # # # # -=O=- # # # # -=O=- # # # # -=O=- # # # # -=O=- # # # # -=O=- # # # # -=O=- # # # #

Info

2021-08-19 13:38:33.376 IST

Watcher not running

Info

2021-08-19 13:38:33.378 IST

Delegate not running

Resolution

Configure the delegate after setting the value for “POLL_FOR_TASKS” to “true” in yaml & change the image type for nodes while creating the cluster.

Diagnostic Steps

  • Configure the Delegate after setting the value for “POLL_FOR_TASKS” to “true” in yaml. After this the container creation step will still fail with the following error message.

Failed to pull image "GCR URL": rpc error: code = NotFound desc = failed to pull and unpack image "GCR URL": failed to unpack image on snapshotter overlayfs: failed to extract layer sha256:eada838fb7777f76b5776c6b4f1cd7694f86a49c39001c931ef04ca324472614: failed to get reader from content store: content digest sha256:9a960853d26c8ea4055375ba6debc7e6436c4ad0c69d97c890209bfb2b6cc212: not found Failed

  • To solve the above usecase change the image type for nodes while creating the cluster with Container-optimized OS with Docker instead of Container-optimized OS with Containerd