Version: Spectra Detect 5.5.1

Installing large file worker group

Installing the tiscale-worker chart in "large mode" will install a tiscale-worker deployment that manages a set of large file processing pods.

A release installed in "large mode" will not install RabbitMQ, PostgreSQL or a shared PersistentVolumeClaim. Instead, large file processing pods will use those resources from another tiscale-worker release (the main release).

Installation

Install Spectra Detect Manager (if it’s not already installed)
Install the main tiscale-worker release with the value worker.largeFileProxy set to true, or upgrade an existing release:

helm upgrade tiscale-worker oci://registry-1.docker.io/appliancesdevs/tiscale-worker \
  --reuse-values \
  --set "worker.largeFileProxy=true" \
  --namespace ns1
kubectl rollout restart deployment tiscale-worker \
  --namespace ns1

Setting worker.largeFileProxy to true enables forwarding of large files to the second tiscale-worker release.

Install tiscale-worker in "large file mode"

Enable large mode (worker.largeFileWorker=true)
It must be installed in the same namespace as the main tiscale-worker release
Set c1000.releaseNamespace to the same value as for the main tiscale-worker release
Disable ingress (ingress.enabled=false) since the main tiscale-release will be used for file ingress.
If the name of the main release is not tiscale-worker, set it as the value worker.mainReleaseName.
Set resources.requests to more than 50% of the resources of large file processing nodes so only one tiscale-worker pod runs on each node.

helm install tiscale-worker-large oci://registry-1.docker.io/appliancesdevs/tiscale-worker \
--set "ingress.enabled=false" \
--set "persistence.storageClassName=efs-sc" \
--set "c1000.releaseNamespace=c1000" \
--set "worker.largeFileWorker=true" \
--set "autoscaling.enabled=true" \
--set "autoscaling.minReplicas=0" \
--set "autoscaling.maxReplicas=3" \
--set "autoscaling.targetCPUUtilizationPercentage=0" \
--set "autoscaling.targetMemoryUtilizationPercentage=0" \
--set "autoscaling.targetInputQueueSize=3" \
--set-json='nodeSelector={"eks.amazonaws.com/nodegroup": "large"}' \
--set-json='tolerations=[{"key": "tiscale-worker/large-file-processing", "operator": "Exists", "effect": "NoSchedule"}]' \
--set-json='resources.requests={"memory": "28Gi", "cpu": "7000m"}' \
--namespace ns1

In the Spectra Detect Manager Central Configuration, under "Worker Configuration -> File Processing", ensure that the Processing Mode is set to "Advanced".

Installation​

Installation