add sdctl procedure in troubleshooting guide

[henderiw]

we install by default for now, optimization can come in the future

[hansthienpondt]

Dependent on https://github.com/sdcio/sdctl/issues/5

[hansthienpondt]

• Make sure schema-server ; data-server and cache are binding their socket on 0.0.0.0 (or ::) and a non-overlapping port (https://github.com/sdcio/config-server/blob/main/artifacts/configmap-data-server.yaml)

[hansthienpondt]

• Make sure schema-server ; data-server and cache are binding their socket on 0.0.0.0 (or ::) and a non-overlapping port (https://github.com/sdcio/config-server/blob/main/artifacts/configmap-data-server.yaml)

This is done by default. time="2024-02-21T15:18:48Z" level=info msg="starting server on :56000"

❯ kubectl exec sdctl -- /app/sdctl -a data-server.network-system.svc.cluster.local:56000 --format json datastore get --ds default.sr1 | jq .target { "type": "netconf", "address": "172.22.1.11:830", "status": 1 }

[henderiw]

so it means we just need to add a deployment for the sdctl pod ?

[hansthienpondt]

so it means we just need to add a deployment for the sdctl pod ?

Correct, This can be a Deployment, a Pod (sleeping for x amount of time with a restart policy), or a Job.

[henderiw]

I prefer a deployment since this allows to upgrade. Can you do a proposal and we can add it to the artefacts

[hansthienpondt]

Cfr, internal discussion, after some testing I would go for:

kubectl run -ti --rm sdctl --image=ghcr.io/sdcio/sdctl:v0.0.1 --restart=Never --command -- /bin/sh

Will check with @steiler how to get this documented!

[hansthienpondt]

Done