must-gather doesn't anonymize many things when running Scylla in GCP

[gdubicki]

What happened?

I did a run of must-gather as documented at https://operator.docs.scylladb.com/stable/support/must-gather.html.

After grepping for some names I noticed that the data collected didn't anonymize some things:

• GCS bucket names used for backups,
• GCP project names,
• GCR image names,

What did you expect to happen?

I was expecting these names to be anonymized.
Instead I had to do a bunch of recursive find and replace (grep -rl old . | xargs sed -i "" -e 's/old/new/g') myself...

At the minimum, the warning in the docs about checking the gathered data should be emphasized and turn into a required step.

How can we reproduce it (as minimally and precisely as possible)?

1. Deploy Scylla in GCP
2. Configure backups in GCS
3. Run some other workloads in the same cluster with images from GCR
4. Run must-gather
5. grep for the aforementioned names in the result directory

Scylla Operator version

1.13.0

Kubernetes platform name and version

$ kubectl version
Client Version: v1.29.6
Kustomize Version: v5.0.4-0.20230601165947-6ce0bf390ce3
Server Version: v1.29.5-gke.1192000

Please attach the must-gather archive.

I can't attach the non-anonymized archive because that's the point here. The additionally manually anonymized version is in #2016.

Anything else we need to know?

No response

[scylla-operator-bot]

The Scylla Operator project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 30d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle stale
• Close this issue with /close
• Offer to help out

/lifecycle stale

[scylla-operator-bot]

The Scylla Operator project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 30d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle rotten
• Close this issue with /close
• Offer to help out

/lifecycle rotten

[gdubicki]

Still valid! I just had to anonymize manually again in #2133.

[gdubicki]

/remove-lifecycle stale

[gdubicki]

/remove-lifecycle rotten

[tnozicka]

After grepping for some names I noticed that the data collected didn't anonymize some things:

GCS bucket names used for backups,
GCP project names,
GCR image names,

None of these are part of the API so I don't think we can automatically redact them and sometimes some of them may matter.

Although, I struggle to see the secrecy of say GCR image names.

At the minimum, the warning in the docs about checking the gathered data should be emphasized and turn into a required step.

It already says you may want to review it. I don't think "required" is fitting here. In Kubernetes secret data is supposed to be stored in Secrets.

By default, all collected Secrets are censored to avoid sending sensitive data. That said, you can always review the archive before you attach it to an issue or your support request.

https://operator.docs.scylladb.com/v1.14/support/must-gather.html#gathering-data-with-must-gather