From 867f268150a657a1c6e21bc4a5199dc497f2d689 Mon Sep 17 00:00:00 2001 From: Maisie Wang Date: Wed, 18 Sep 2019 11:06:46 -0700 Subject: [PATCH 1/2] Add setup job troubleshooting steps --- deploy/README.md | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/deploy/README.md b/deploy/README.md index 16b52f3085..9853f7009a 100644 --- a/deploy/README.md +++ b/deploy/README.md @@ -716,6 +716,25 @@ USER fluent # Troubleshooting Collection +## `helm install` hanging + +If `helm install` hangs, it usually means the pre-install setup job is failing and is in a retry loop. Due to limitation of helm, errors from the setup job cannot be fed back to the `helm install` command. Kubernetes schedules the job in a pod, so you can look at logs from the pod to see why the job is failing. First find the pod name in the namespace where the helm chart is deployed: +```sh +kubectl get pods -n sumologic +``` + +Get the logs from that pod: +``` +kubectl logs POD_NAME -f +``` + +If you see `Secret 'sumologic::sumologic' exists, abort.` from the logs, delete the existing secret: +``` +kubectl delete secret sumologic -n sumologic +``` + +`helm install` should proceed after the existing secret is deleted before exhausting retries. If it did time out after exhausting retries, rerun the `helm install` command. + ## Namespace configuration The following `kubectl` commands assume you are in the correct namespace `sumologic`. By default, these commands will use the namespace `default`. From 0dda47df13809747c8ebd24e9ce92f4c00cddbd8 Mon Sep 17 00:00:00 2001 From: David Duke Date: Wed, 18 Sep 2019 12:01:28 -0700 Subject: [PATCH 2/2] Update README --- deploy/README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/deploy/README.md b/deploy/README.md index 9853f7009a..4750f03536 100644 --- a/deploy/README.md +++ b/deploy/README.md @@ -718,7 +718,7 @@ USER fluent ## `helm install` hanging -If `helm install` hangs, it usually means the pre-install setup job is failing and is in a retry loop. Due to limitation of helm, errors from the setup job cannot be fed back to the `helm install` command. Kubernetes schedules the job in a pod, so you can look at logs from the pod to see why the job is failing. First find the pod name in the namespace where the helm chart is deployed: +If `helm install` hangs, it usually means the pre-install setup job is failing and is in a retry loop. Due to a Helm limitation, errors from the setup job cannot be fed back to the `helm install` command. Kubernetes schedules the job in a pod, so you can look at logs from the pod to see why the job is failing. First find the pod name in the namespace where the Helm chart is deployed: ```sh kubectl get pods -n sumologic ```