Open eyanez111 opened 2 years ago
Hi @eyanez111
In your example config above you are saying that the cert is at /etc/ca.crt, so you would mount the secret as a volume at /etc/ca.crt. Take a look at this example, the person who needed the cert was using it to connect to Elasticsearch, but it does not matter. Just create the secret, create the volume mount, and create the volume. I have not added a cert to Heartbeat this way, but I have done this for other software.
Thanks for the answer. I really appreciate it.
So this is the example provided by Elastic:
ssl:
certificate_authorities: ['/etc/ca.crt']
supported_protocols: ["TLSv1.0", "TLSv1.1", "TLSv1.2"]
I assume then that based on the link you shared if I already have the cert on an existing secret that the website uses all what I have to do is:
ssl:
certificate_authorities: ['/etc/ca.crt']
supported_protocols: ["TLSv1.0", "TLSv1.1", "TLSv1.2"]
volumeMounts:
- name: cert
mountPath: /etc/ssl/certs
...
volumes:
- name: cert
secret:
secretName: **cert-from-the-website-I-want-to-monitor**
defaultMode: 0400
or I do not need the part: ssl: certificate_authorities: ['/etc/ca.crt'] supported_protocols: ["TLSv1.0", "TLSv1.1", "TLSv1.2"]
Thanks again
I think you have it mostly correct, just change this piece:
volumeMounts:
- name: cert
mountPath: /etc/ssl/certs
to
volumeMounts:
- name: cert
mountPath: /etc/ca.crt
subPath: config.json
readOnly: true
If it does not work, then have a look at the logs and shell into the container and have a look at the file mounted at /etc/ca.crt
. It has been a couple years since I worked on Beats in k8s.
There is also an example that I wrote years ago that shows how to create a configmap that contains the cert and mount it. The cert in the example is the cert to connect to Elasticsearch, so that is not what you want (I think in your use case you want to give Heartbeat a cert to connect to a service that you are managing, correct?), but the technique is the same.
Hello @DanRoscigno,
Sorry it did not work I did:
ssl:
certificate_authorities: ['/etc/ca.crt']
supported_protocols: ["TLSv1.0", "TLSv1.1", "TLSv1.2"]
volumeMounts:
- name: cert
mountPath: /etc/ca.crt
subPath: config.json
readOnly: true
...
volumes:
- name: cert
secret:
secretName: **cert-from-the-website-I-want-to-monitor**
defaultMode: 0400
What goes on "subPath: config.json" do I need to leave it like that or should I actually put a path?
the name of the file that you want is subPath (ca.crt). I find it confusing too.
thanks for answering so fast. So it should be like:
volumeMounts:
name: cert mountPath: /etc/ca.crt subPath: config.json ---> name of the path for the ca.crt? readOnly: true
but how do I point at any path if the cert I have it on a secret?
thanks
three steps:
Create secret
Make a volume
named cert
Mount the volume
certin a
volumeMountalso named
cert. This third step is where you specify the filename, by specifying both the
mountPathand
subPath. The
mountPathincludes the dir name and filename, and the
subPath` is just the filename.
One question. I do have the secret but is in a different namespace do I have to copy that secret and put it on the same namespace that I am running it? I think that is the main thing why it is failing.
as per what you said then I think is like this right
ssl:
certificate_authorities: ['/etc/**SECRETNAME**']
supported_protocols: ["TLSv1.0", "TLSv1.1", "TLSv1.2"]
volumeMounts:
- name: cert
mountPath: /etc/**SECRETNAME**
subPath: **SECRETNAME**
readOnly: true
volumes:
- name: cert
secret:
secretName: **SECRETNAME**
defaultMode: 0400
Thanks a lot for all the help
My pleasure!
The secret definitely needs to be in the same namespace unless your pod is running with admin level.
kubectl describe pod and kubectl exec -it demo-pod -- /bin/sh will help you debug it. Open a shell into the pod and look at /etc/ca.crt to see if it contains the right thing, and describe will tell you about the secret and volume.
ah sorry I just talked to DEV and looks like we do not have a hold of the cert. We only have the tls.cert. So there is not a CA involved. On the secret I just have the tls.cert and the tls.key. Is it possible to do it like that:
ssl:
certificate_authorities: ['/etc/tls.crt']
supported_protocols: ["TLSv1.0", "TLSv1.1", "TLSv1.2"]
volumeMounts:
- name: cert
mountPath: /etc/tls.crt
subPath: tls.crt
readOnly: true
volumes:
- name: cert
secret: kubernetes.io/tls
secretName: CERT-NAME
defaultMode: 0400
again deeply thank you as not even Elastic support has been able to point me the right direction.
I would get it working without k8s first, then work on k8s.
sorry but I just have a k8s environment.
ah, I was able to get the ca after talking to devs and IT. So I was ready to proceed but after:
volumeMounts:
- name: cert
mountPath: /etc/tls.cert
readOnly: true
subPath: tls.crt
volumes:
- name: cert
secret: kubernetes.io/tls
secretName: NAME
defaultMode: 0400
This is what I get:
error: error validating "heartbeat-kubernetes-prod-rancher.yaml": error validating data: [ValidationError(Deployment.spec.template.spec.volumes[0]): unknown field "defaultMode" in io.k8s.api.core.v1.Volume, ValidationError(Deployment.spec.template.spec.volumes[0].secret): invalid type for io.k8s.api.core.v1.SecretVolumeSource: got "string", expected "map", ValidationError(Deployment.spec.template.spec.volumes[0]): unknown field "secretName" in io.k8s.api.core.v1.Volume]; if you choose to ignore these errors, turn validation off with --validate=false
so not sure why is not recognizing the secretName... any idea? I think I am very very close!
I don't think secretName
belongs in the cert volume. Look at this example, and setup the cert like that. If you need more details on the volumes, etc., have a look at the k8s docs.
I found the error. I was adding what type of secret and secret is just to define that it is a secret what is trying to mount. so I just left it as Secret that worked. I see now the cert in the pod. I think I am quite close.... thanks again for all the help without your assistance I would not have been able to do this.
my pleasure, once you get it all working can you pop the manifest in here for the next person?
for sure, I am still working on it but at least I see the mount and the cert there. So definitely heading the right direction
I have not been able to make it work with the tls but I think this means it will only work with the ca.crt and not with the tls.crt. Still the process what you described @DanRoscigno it was pretty much this:
apiVersion: v1
kind: Secret
metadata:
name: self-signed-certificate-secret
data:
ca-certificates.crt: <base 64 encoded string of certificate>
volumes:
-
secret:
secretName: NAME FOR YOUR SECRET
defaultMode: 0400
name: cert
Mount the volume certin avolumeMountalso namedcert. This third step is where you specify the filename, by specifying both the mountPathandsubPath. The mountPathincludes the dir name and filename, and thesubPath` is just the filename.
volumeMounts:
-
mountPath: /etc/ca.crt
name: cert
readOnly: true
subPath: ca.crt
- type: http
id: service
name: NAME
hosts: ["https://YOUR-URL"]
schedule: '@every 5s'
ssl:
certificate_authorities: ['/etc/ca.crt']
supported_protocols: ["TLSv1.0", "TLSv1.1", "TLSv1.2"]
That is how you add a secret based on a CA to heartbeat thanks again @DanRoscigno
Hello Community,
https://www.elastic.co/guide/en/beats/heartbeat/7.16/monitor-http-options.html#monitor-http-tls-ssl
What I do not find is how to add those certs to the pod? I have the secrets of the certs but I am not sure if I should create a secret for heatbeat? or if I should modify the existing one that I see in the kube-system:
heartbeat-token-lk8sl kubernetes.io/service-account-token 3 2d1h
if I have to create one do you have a process on how by creation it will be called by hearbeat? do I have to use any apiVersion? and if it is done by modifying the heartbeat token do I need to replace the ca.crt that I see there?
Thanks eyanez111