Blockstor migration guide scales piraeus-operator in a fixed namespace and never scales it back #813

content/en/docs/next/storage/blockstor-migration.md (branch docs/blockstor-storage-backend, #667) scales the operator down with kubectl -n cozy-system scale deploy/piraeus-operator-controller-manager --replicas=0 (line 40).

On some installations the operator runs in cozy-linstor, not cozy-system, so the command fails and the operator keeps running and rebuilding what the next steps scale down.

The guide then says the operator stays on in external mode to keep driving the CSI (line 128), but no step scales it back to 1. The order matters: brought back before adoption, it starts linstor-csi-controller again and undoes the CSI freeze the guide asks for at step 6.

The guide also does not cover a cluster where the piraeus-operator HelmRelease was suspended by hand. A suspended release receives none of the chart changes the switch relies on.

Proposal: find the operator's namespace instead of assuming it, add a step that scales it back after adoption and says why not earlier, and check for a suspended HelmRelease before starting.