CockroachDB - the open source, cloud-native distributed SQL database.
Certificate Self-Signer utility is developed to allow the cockroachdb helm chart to be able to deploy secure cluster, without any dependency on the outside tool to create or sign its certificate.
You can enable/disable this utility by setting the tls.certs.selfSigner.enabled
option as true/false.
This option allow cockroachdb to generate the CA, node and client certificates and use those certificates to form a secure cockroachdb cluster. User can configure the duration and expiry window of each certificate types. Following are the options provided as default values in hours.
# Minimum Certificate duration for all the certificates, all certs duration will be validated against this.
tls.certs.selfSigner.minimumCertDuration: 624h
# Duration of CA certificates in hour
tls.certs.selfSigner.caCertDuration: 43800h
# Expiry window of CA certificates means a window before actual expiry in which CA certs should be rotated.
tls.certs.selfSigner.caCertExpiryWindow: 648h
# Duration of Client certificates in hour
tls.certs.selfSigner.clientCertDuration: 672h
# Expiry window of client certificates means a window before actual expiry in which client certs should be rotated.
tls.certs.selfSigner.clientCertExpiryWindow: 48h
# Duration of node certificates in hour
tls.certs.selfSigner.nodeCertDuration: 8760h
# Expiry window of node certificates means a window before actual expiry in which node certs should be rotated.
tls.certs.selfSigner.nodeCertExpiryWindow: 168h
These durations can be configured by user with following validations:
- CaCertExpiryWindow should be be greater than minimumCertDuration.
- Other certificateDuration - certificateExpiryWindow should be greater than minimumCertDuration.
This utility also handles certificate rotation when they come near expiry. You can enable or disable the certificate rotation with following setting:
# If set, the cockroachdb cert selfSigner will rotate the certificates before expiry.
tls.certs.selfSigner.rotateCerts: true
If user has a custom CA which they already use for certificate signing in their organisation, this utility provides a way for user to provide the custom CA. All the node and client certificates are signed by this user provided CA.
To provide the CA certificate to the crdb you have to create a tls certificate with ca.crt
and ca.key
and provide the
secret as:
# If set, the user should provide the CA certificate to sign other certificates.
tls.certs.selfSigner.caProvided: true
# It holds the name of the secret with caCerts. If caProvided is set, this can not be empty.
tls.certs.selfSigner.caSecret: "custom-ca-secret"
You will still have options to configure the duration and expiry window of the certificates:
# Minimum Certificate duration for all the certificates, all certs duration will be validated against this.
tls.certs.selfSigner.minimumCertDuration: 624h
# Expiry window of CA certificates means a window before actual expiry in which CA certs should be rotated.
tls.certs.selfSigner.caCertExpiryWindow: 648h
# Duration of Client certificates in hour
tls.certs.selfSigner.clientCertDuration: 672h
# Expiry window of client certificates means a window before actual expiry in which client certs should be rotated.
tls.certs.selfSigner.clientCertExpiryWindow: 48h
# Duration of node certificates in hour
tls.certs.selfSigner.nodeCertDuration: 8760h
# Expiry window of node certificates means a window before actual expiry in which node certs should be rotated.
tls.certs.selfSigner.nodeCertExpiryWindow: 168h
This utility will only handle the rotation of client and node certificates, the rotation of custom CA should be done by user.
When user install cockroachdb cluster with self-signer enabled, you will see the self-signer job.
kubectl get pods
NAME READY STATUS RESTARTS AGE
crdb-cockroachdb-self-signer-mmxp8 1/1 Running 0 15s
This job will generate CA, client and node certificates based on the user input mentioned in previous section. You can see the following secrets representing each certificates:
kubectl get secrets
NAME TYPE DATA AGE
crdb-cockroachdb-ca-secret Opaque 2 3m10s
crdb-cockroachdb-client-secret kubernetes.io/tls 3 3m9s
crdb-cockroachdb-node-secret kubernetes.io/tls 3 3m10s
crdb-cockroachdb-self-signer-token-qcc72 kubernetes.io/service-account-token 3 3m29s
crdb-cockroachdb-token-jpbms kubernetes.io/service-account-token 3 3m8s
default-token-gmhdf kubernetes.io/service-account-token 3 11m
sh.helm.release.v1.crdb.v1 helm.sh/release.v1 1 3m30s
After this, the cockroachdb init jobs starts and copies this certificate to each nodes:
prafull@EMPID18004:helm-charts$ kubectl get pods
NAME READY STATUS RESTARTS AGE
crdb-cockroachdb-0 0/1 Init:0/1 0 18s
crdb-cockroachdb-1 0/1 Init:0/1 0 18s
crdb-cockroachdb-2 0/1 Init:0/1 0 18s
crdb-cockroachdb-init-fclbb 1/1 Running 0 16s
At last, the cockroach db cluster comes into running state with following output:
helm install crdb ./cockroachdb/
NAME: crdb
LAST DEPLOYED: Thu Aug 19 18:03:37 2021
NAMESPACE: crdb
STATUS: deployed
REVISION: 1
NOTES:
CockroachDB can be accessed via port 26257 at the
following DNS name from within your cluster:
crdb-cockroachdb-public.crdb.svc.cluster.local
Because CockroachDB supports the PostgreSQL wire protocol, you can connect to
the cluster using any available PostgreSQL client.
Note that because the cluster is running in secure mode, any client application
that you attempt to connect will either need to have a valid client certificate
or a valid username and password.
Finally, to open up the CockroachDB admin UI, you can port-forward from your
local machine into one of the instances in the cluster:
kubectl port-forward crdb-cockroachdb-0 8080
Then you can access the admin UI at https://localhost:8080/ in your web browser.
For more information on using CockroachDB, please see the project's docs at:
https://www.cockroachlabs.com/docs/
Kick off the upgrade process by changing the new Docker image, where $new_version
is the CockroachDB version to which you are upgrading:
helm upgrade my-release cockroachdb/cockroachdb \
--set image.tag=$new_version \
--reuse-values
Kubernetes will carry out a safe rolling upgrade of your CockroachDB nodes one-by-one. Monitor the cluster's pods until all have been successfully restarted:
Kubernetes signed certificates is deprecated from the Kubernetes v1.22+ and user will not be able to use this methods for signing certificates.
User can move from old kubernetes signing certificates by performing following steps:
Run the upgrade command with upgrade strategy set as "onDelete" which only upgrades the pods when deleted by the user.
helm upgrade crdb-test cockroachdb --set statefulset.updateStrategy.type="OnDelete"
While monitor all the pods, once the init-job is created, you can delete all the cockroachdb pods with following command:
kubectl delete pods -l app.kubernetes.io/component=cockroachdb
This will delete all the cockroachdb pods and restart the cluster with new certificates generated by the self-signer utility. The migration will have some downtime as all the pods are upgraded at the same time instead of rolling update.