Open Halogenmake opened 4 months ago
Hi @Halogenmake. Thanks for the question.
@kiblik Thank you for your response! I don't see the job being launched. It's not in the namespace where I'm trying to install DefectDojo (namespace: defectdojo) nor in the default.
There is ttlSecondsAfterFinished
for this job so there is a chance that it was already removed.
Can you trigger an upgrade of the chart and check how is initializer
behaving?
If it is not triggered, double-check that .Values.initializer.run
is true.
It seems that the initializer is not triggering and not starting, even though its parameter is set to true.
initializer:
run: true
jobAnnotations: {helm.sh/hook: "post-install,post-upgrade"}
annotations: {}
labels: {}
keepSeconds: 60
affinity: {}
nodeSelector: {}
resources:
requests:
cpu: 100m
memory: 256Mi
limits:
cpu: 2000m
memory: 512Mi
I can only recommend running helm template --debug
/helm install --dry-run --debug
/--dry-run=server
, check if initializer is defined there and dig deeper there. I do not have any other recommendation here.
But please provide feedback if you will find out what was the reason. I'm quite curious.
@kiblik It's a very strange situation. If I install using Helm from the downloaded repository, the job starts and everything is fine, but if I install through Terragrunt, the job refuses to start. When installing through Terragrunt, I specify the following repository:
resource "helm_release" "defectdojo" {
name = "defectdojo"
repository = "https://raw.githubusercontent.com/DefectDojo/django-DefectDojo/helm-charts"
chart = "defectdojo"
namespace = var.namespace
create_namespace = true
timeout = 600
version = var.chart_version
values = [
var.extra_values
]
@kiblik It's a very strange situation. If I install using Helm from the downloaded repository, the job starts and everything is fine, but if I install through Terragrunt, the job refuses to start. When installing through Terragrunt, I specify the following repository:
resource "helm_release" "defectdojo" { name = "defectdojo" repository = "https://raw.githubusercontent.com/DefectDojo/django-DefectDojo/helm-charts" chart = "defectdojo" namespace = var.namespace create_namespace = true timeout = 600 version = var.chart_version values = [ var.extra_values ]
@Halogenmake, I have zero experience with Terragrunt. I'm not able to help you here. If you would find a fix, feel free to share.
I have the same issue and it's quite strange that helm command deploys initializer container but terraform helm_Release resource doesn't. Both manual command and terraform use same values.yaml with initializer enabled.
If someone else will have same issue with terraform just add the following parameter to the helm_release resource to fix initializer:
wait = false # Fix initializer
Can we rename this issue? It is still not solved but we identified that the issue is with "not triggered initializer in terraform deployment". Or maybe even close it because the issue is outside of DefectDojo.
Bug description Good day! I'm trying to install DefectDojo on my k8s cluster. We use Terragrunt for the installation (essentially the same as Helm). During the installation process, the defectdojo-django-*- pod does not deploy. In the uWSGI image, the following error occurs:
django.db.utils.ProgrammingError: (1146, "Table 'defectdojo.django_site' doesn't exist").
For my installation, I use a MySQL database. When I connect to the database locally, I indeed do not see this table.
I am installing the latest release from the master branch. The version of the HELM chart is 1.6.129. Can you advise how to solve this problem?
The installation is performed with the following parameters, different from the default settings:
Deployment method (select with an
X
)Environment information
Logs
Screenshots