-
Notifications
You must be signed in to change notification settings - Fork 16.8k
Grafana helm chart pod failing #4267
Comments
I've submitted a PR with the same strategy I used with the rabbitmq-ha chart. It still feels a little clunky to do this, but unless containers stop writing to their configdirs or Kubernetes allows writing to ConfigMaps again (or adds some primitives to do what I did in this PR automatically), I don't see any other way. |
For a temporary workaround, you can add the following command for the docker container in the deployment.yaml (taken from the init script) command: ["/bin/bash", "-c", "chmod 777 /var/lib/grafana/ && gosu grafana /usr/sbin/grafana-server --homepath=/usr/share/grafana --config=/etc/grafana/grafana.ini cfg:default.paths.provisioning=/etc/grafana/provisioning cfg:default.log.mode=console cfg:default.paths.data=/var/lib/grafana/ cfg:default.paths.logs=/var/lib/grafana cfg:default.paths.plugins=/var/lib/grafana/plugins"] |
Related: grafana/grafana-docker#141 |
but looks like the fix is on its way: |
alternative workaround: in server object: extraEnv:
- name: GF_PATHS_CONFIG
value: /etc/grafana-cm/grafana.ini
configLocalPath: /etc/grafana-cm
dashboardLocalPath: /var/lib/grafana-dashboards and modify grafana.ini cm: [dashboards.json]
enabled = true
path = /var/lib/grafana-dashboards edit: i hate md |
Thanks @csabakollar and all for the tips. I've been banging my head with this for a few hours. I was able to just switch my image to |
…figMap to an EmptyDir (#4271) * Update deployment.yaml ConfigMaps are mounted read-only since Kubernetes 1.9.4 (kubernetes/kubernetes#58720). The Grafana Chart uses a ConfigMap to provision the config- and dashboard directories. Grafana tries to create/modify files in these directories, which is not allowed anymore. This PR adds an busybox initContainer to the Deployment that copies the files from the ConfigMap to a new emptyDir, similar to #4169. Fixes #4267. * bump Chart version
…figMap to an EmptyDir (helm#4271) * Update deployment.yaml ConfigMaps are mounted read-only since Kubernetes 1.9.4 (kubernetes/kubernetes#58720). The Grafana Chart uses a ConfigMap to provision the config- and dashboard directories. Grafana tries to create/modify files in these directories, which is not allowed anymore. This PR adds an busybox initContainer to the Deployment that copies the files from the ConfigMap to a new emptyDir, similar to helm#4169. Fixes helm#4267. * bump Chart version
…figMap to an EmptyDir (helm#4271) * Update deployment.yaml ConfigMaps are mounted read-only since Kubernetes 1.9.4 (kubernetes/kubernetes#58720). The Grafana Chart uses a ConfigMap to provision the config- and dashboard directories. Grafana tries to create/modify files in these directories, which is not allowed anymore. This PR adds an busybox initContainer to the Deployment that copies the files from the ConfigMap to a new emptyDir, similar to helm#4169. Fixes helm#4267. * bump Chart version
…figMap to an EmptyDir (helm#4271) * Update deployment.yaml ConfigMaps are mounted read-only since Kubernetes 1.9.4 (kubernetes/kubernetes#58720). The Grafana Chart uses a ConfigMap to provision the config- and dashboard directories. Grafana tries to create/modify files in these directories, which is not allowed anymore. This PR adds an busybox initContainer to the Deployment that copies the files from the ConfigMap to a new emptyDir, similar to helm#4169. Fixes helm#4267. * bump Chart version Signed-off-by: voron <av@arilot.com>
Is this a BUG REPORT or FEATURE REQUEST?: Bug Report
Version of Helm and Kubernetes:
Helm version: 2.8.2
Kuernetes Version: 1.9.4
Which chart: stable/grafana
What happened: grafana pod fails to start with the following error.This was previously working, and discussion in the rook.io slack with @galexrt turned up that this may be related to a recent security patch in kubernetes.
How to reproduce: Install the helm chart with default configurations on a node running 1.9.4
The text was updated successfully, but these errors were encountered: