I do have a GitRepo with 3 paths, 1*helmchart, 2*kustomize. Is it normal that fleet creates a helmchart (or at least a helm-secret) for each of the kustomize-folders?

Hi, I'm unable to edit a GitRepo configuration because the resource version continuously changes. Have you experienced this? How can I edit the GitRepo without recreate it?

Hello everybody, I'm contacting you because I'm trying to set up Keda and as usual I want to use Fleet to deploy its objects. Keda is installed (still KubeVersion problem but that's not the topic here. I deployed a fleet that contains a helm chart of a project of mine with, among other things, two deployments and 2 ScaledObject. Just like the HPAs, Keda modifies the replicas so I added a diff in fleet and it works fine. On the other hand, ScaledObject objects have an “Activating” status when the trigger is pending. This is a normal status (from what I understand). But Rancher sees it in error on its UI and the bundle is also in error.

NotReady(1) [Bundle ebad]; scaledobject.keda.sh ebad-ma7/ebad-back error] Scaling is not performed because triggers are not active, Resource is Ready; scaledobject.keda.sh ebad-ma7/ebad-back modified {"spec":{"triggers":[{"metadata":{"desiredReplicas":"2","end":"35 14 * * *","start":"29 14 * * *","timezone":"Europe/Paris"},"type":"cron"}]}}; scaledobject.keda.sh ebad-ma7/ebad-front error] Scaling is not performed because triggers are not active, Resource is Ready

Is there a way to ignore this type of status? I think it's related to this issue: https://github.com/rancher/fleet/issues/937

is there a way to patch resource coming from a helm chart with kustomize?

Question... it has occurred now many times that resources delete (on sync) and cause rewrite of resources?? causes some downtime, it is also possible that on sync resources are not available and hence it happens, is there any threshold?

Hi folks! I was reading over the Fleet documentation regarding how a

GitRepo

is mapped to clusters, as well as the

ClusterGroup

concepts, and they're both using label selectors if I'm not mistaken. I'm wondering - is there a doc or proposal for how Fleet manages the case where the set of clusters matching the label selector changes? In "normal" kube, e.g. a

NodeSelector

will allow for transparent selection, movement, re-scheduling, etc - is that also what happens in Fleet?

it syncs indeed, occasionally

Sorry, not sure my question was phrased well - for instance, if I have some

GitRepo

that should select

region=us-west,cloud=aws,machine-type=xxl

or whatever, will I get instant failover if a cluster in that

ClusterGroup

goes down? Is there anything that balances out the usage of the group so, as an admin providing these clusters, I have less maintenance to do?

Has anyone used fleet to deploy the Rancher applications (https://git.rancher.io/charts) across multiple downstream clusters? I'm particularly interested in trying to roll out Rancher CIS Benchmarking.

I am working on a rancher implementation v2.69 with fleet. In the gitrepo object I have it pointing to a cluster group. In the git repo I have a few bundles. What I am trying to do is have a bundle only applied to a certain cluster. The documentation is not clear on how this is done. I could create multiple git repo of the same git project and define it differently but I was wondering if there is a way I could further enhance the filter through fleet.yaml in the bundle that it resides in

HI! I need to force pause the fleet git update process at a scheduled time. How can I schedule this operation? Thank you.

You may need to do it repo by repo (there's no global flag to pause all syncs AFAIK). If you go to your git repo object you can set the pause attribute to true. Example: apiVersion: fleet.cattle.io/v1alpha1 kind: GitRepo metadata: name: fleet-examples-single-cluster-helm annotations: {} # key: string creationTimestamp: '2022-10-02T21:59:18Z' generation: 4 labels: {} # key: string namespace: fleet-default resourceVersion: '85196906' uid: a2cd4b38-fc81-474b-bbd1-0ac0a5d938ad fields: - fleet-examples-single-cluster-helm - https://github.com/juanbrny/fleet-examples.git - 8b480f244d0b4edb51cdbb048dc8aa5488e53f66 - 0/0 - null spec: branch: master insecureSkipTLSVerify: false paths: - /single-cluster/helm # - string paused: true repo: https://github.com/juanbrny/fleet-examples.git targets: - clusterSelector: matchExpressions: - key: provider.cattle.io operator: NotIn values: - harvester

May that work for you @numerous-lighter-90852?

Instead of using Kubernetes API to change that you can also use Rancher's REST API. Visit: https://myclusterfqdn/v1/fleet.cattle.io.gitrepos/fleet-default/mygit-repo-name and click on Edit to see an example request you can just use with curl to modify the paused attribute.

Thanks @quick-sandwich-76600, I test today this tips and I'll let you know.

Hi @quick-sandwich-76600. Thanks to your idea, I came up with this solution:

kubectl patch gitrepo/<repo name> -n fleet-local -p '{"spec": {"paused":false}}' --type merge

What do you think about it?

Yes, that should work great @numerous-lighter-90852 👍 At the end of the day, all Fleet and Rancher objects are just CRDs that can be easily modified/automated either through the standard Kubernetes way or through REST API (that's why I added the other example).

Thank you, it is clearer to me now. TOP! 😄

I really like Fleet; it is young, but it works well. I have joined the SUSE community.

👋 … how can you handle CRD’s in fleet? I have applications that is part helm chart and part CRD’s that are usually applied with kubectl BUT can I do that with fleet? ex. for appscode Stash

# Update catalog CRDs
$ kubectl apply -f <https://github.com/stashed/installer/raw/v2023.03.13/crds/stash-catalog-crds.yaml>

# Update the helm repositories
$ helm repo update

# Upgrade Stash Community operator chart
$ helm upgrade stash appscode/stash \
  --version v2023.03.13 \
  --namespace kube-system \
  --set features.community=true               \
  --set-file global.license=/path/to/the/license.txt

Hello everyone, I have an issue with Fleet and the Rancher CD Feature. Rancher does not seem to overwrite the version in targetCustomizations. Here is an example fleet.yaml:

namespace: example-namespace

helm:

releaseName: dfl

chart: "<oci://custom.registry.net/test/example>"

repo: ""

version: "1.0.0"

targetCustomizations:

- name: dev

helm:

chart: "<oci://custom.registry.net/test/example>"

version: "0.0.0-dev"

clusterName: cluster-dev

On our cluster "cluster-dev" it tries to deploy the Version 1.0.0 instead of 0.0.0-dev. Do you know what could be wrong? Thank you in advance.

Hi everyone! I have a weird issue regarding fleet + overlays (kustomize as well as simple manifests; rancher-2.7.1). I am trying to deploy an app, with a cluster-specific configMap:

targetCustomizations:

- name: cluster-0530

clusterName: cluster-0530

yaml:

overlays:

- cluster-0530

- name: cluster-0531

clusterName: cluster-0531

yaml:

overlays:

- cluster-0531

- name: cluster-0532

clusterName: cluster-0532

yaml:

overlays:

- cluster-0532

Now i put the configMap in each cluster-folder, but it will get deployed only on 1 cluster.. Even though

fleet test -t cluster-xyz .

shows the correct manifests.

I created an example - you can see the overlays, but only one of them gets deployed.. https://gitlab.com/strowi/fleet-test/-/tree/master/test

I am using rancher fleet. I wanna know if the fleet agents can communicate with teh fleet manager without needed the certificate (API_SERVER_CA_DATA) for the kube api server. I am essentialy looking to have some auth server in between

@flat-whale-67864 has left the channel

Hi, I have a problem with the rancher-logging-root-fluentbit daemon. Some pods start correctly, others have this error. Error: Error response from daemon: Duplicate mount point: /var/lib/docker/containers Any ideas to solve this error? Thanks.

I have a bundle that needs to be customized on multiple axes (region, environment), I think I'm missing something in how targetCustomization or how clusterSelector work to cover such a use case.