Before upgrading your CDS Instance:
PostgreSQL 14 is now recommended
This release introduced the CDN Artifacts. This means that all artifacts upload / download will be done through the CDN service. This is not enabled by default, you have to enable that with a feature flipping. The version 0.51 will use the CDN service to manage artifacts by default.
Some configuration have to be added to the cdn service to manage artifacts.
You have to add a storageUnits.buffers with the type “file”.
[cdn.storageUnits.buffers.local-buffer]
# it can be 'log' to receive logs or 'file' to receive artifacts
bufferType = "file"
[cdn.storageUnits.buffers.local-buffer.local]
path = "/var/lib/cds-engine/cdn-buffer"
To multi-instantiate the cdn service, you can use a NFS for the bufferType file, example:
[cdn.storageUnits.buffers.buffer-nfs]
bufferType = "file"
[cdn.storageUnits.buffers.buffer-nfs.nfs]
host = "w.x.y.z"
targetPartition = "/zpool-partition/cdn"
userID = 0
groupID = 0
[[cdn.storageUnits.buffers.buffer-nfs.nfs.encryption]]
Cipher = "aes-gcm"
Identifier = "nfs-buffer-id"
## enter a key here, 32 length
Key = "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
Sealed = false
Make sure you have configured cdn in the paragraph above, then enable it with this command:
cat <<EOF > feature.yaml
name: cdn-artifact
rule: return true
EOF
cdsctl admin feature import feature.yaml
In the next release (0.51.0), you don’t need to use the feature flipping cdn-artifact to enable cdn artifact.
If you want to migrate some artifacts to the new backend, you can use cdsctl:
$ cdsctl workflow artifact cdn-migrate <project_key> <workflow_name> <run_number>
This command will migrate artifacts for one workflow run.
The default retention rule have to be added in api configuration:
[api.workflow]
# Default rule for workflow run retention policy, this rule can be overridden on each workflow.
# Example: 'return run_days_before < 365' keeps runs for one year.
defaultRetentionPolicy = "return run_days_before < 365"
Documentation: https://ovh.github.io/cds/docs/concepts/workflow/retention/
A new hatchery configuration attribute is available to control the maximum attempts to start a same job.
Example on the hatchery.local:
[hatchery.local.commonConfiguration.provision]
# Maximum attempts to start a same job. -1 to disable failing jobs when too many attempts
# maxAttemptsNumberBeforeFailure = 5
downloadFromGitHub and supportedOSArch are added. This will allow you to download cds workers / cdsctl from GitHub if it’s not in already downloaded into the directory.
[api.download]
# this directory contains cds binaries. If it's empty, cds will download binaries from GitHub (property downloadFromGitHub) or from an artifactory instance (property artifactory) to it
directory = "/var/lib/cds-engine"
# allow downloading binaries from GitHub
downloadFromGitHub = true
# example: ["darwin/amd64","darwin/arm64","linux/amd64","windows/amd64"]. If empty, all os / arch are supported: windows,darwin,linux,freebsd,openbsd and amd64,arm,386,arm64,ppc64le
supportedOSArch = []