Repository navigation
Conversation
Contributor
|
Commit: Regression test reportSelected 20 changed tests (limit=20, 27 not selected) 38 tests: regression=36 unreleased=0 coverage=2 failing=0
TestAccept/bundle/resources/catalogs/rename/DMS=: branch ✅ | main (6a8d9bf) ❌ | latest (v1.25.0) ❌main (6a8d9bf): latest (v1.25.0): TestAccept/bundle/resources/catalogs/rename/DMS=true: branch ✅ | main (6a8d9bf) ❌ | latest (v1.25.0) ❌main (6a8d9bf): latest (v1.25.0): TestAccept/bundle/resources/catalogs/rename_all_out_of_band/DMS=: branch ✅ | main (6a8d9bf) ❌ | latest (v1.25.0) ❌main (6a8d9bf): latest (v1.25.0): 33 more failure details omitted, see the full report in the run. |
Collaborator
✅ Integration test reportCommit: a44c3dc
Top 37 slowest tests (at least 2 minutes):
|
denik
force-pushed
the
denik/models-rename
branch
from
November 5, 2026 15:30
f364f17 to
18077b3
Compare
denik
force-pushed
the
denik/models-rename
branch
7 times, most recently
from
November 9, 2026 13:09
64af993 to
7496114
Compare
denik
force-pushed
the
denik/models-rename
branch
from
November 9, 2026 14:52
7496114 to
132a53d
Compare
denik
force-pushed
the
denik/models-rename
branch
from
November 9, 2026 14:59
132a53d to
b5f68ef
Compare
denik
force-pushed
the
denik/models-rename
branch
4 times, most recently
from
November 10, 2026 13:23
b0476cd to
36755c1
Compare
Changing a schema's name used to delete and recreate it. It is now an updatable ID field renamed via PATCH new_name, which keeps the schema and everything in it. UC moves the children of a renamed catalog or schema with it. A child volume (or a schema in a renamed catalog) whose catalog_name/schema_name references the parent's name was planned as a recreate, and its create then failed with RESOURCE_ALREADY_EXISTS on every deploy. The planner now re-keys such children (update_id, reason parent_renamed), and they are addressed under their new parents. The test server now moves children on catalog/schema rename and rejects duplicate volume names, matching UC. Co-authored-by: Isaac <no-reply@databricks.com>
Co-authored-by: Isaac <no-reply@databricks.com>
A captured terraform fixture (migrate/command/basic) deploys the same volume name from two targets. The plan goldens already catch a regression to recreate, so the duplicate check is not needed. Co-authored-by: Isaac <no-reply@databricks.com>
Co-authored-by: Isaac <no-reply@databricks.com>
Also restores the grants/schemas/change_name golden for the in-place rename, which the rebase onto denik/grants-rekey dropped. Co-authored-by: Isaac <no-reply@databricks.com>
Also restores the grants/schemas/change_name golden for the in-place rename. Co-authored-by: Isaac <no-reply@databricks.com>
schema-rename now renames in place, so it no longer covers the schema recreate. Co-authored-by: Isaac <no-reply@databricks.com>
Co-authored-by: Isaac <no-reply@databricks.com>
…leReason Co-authored-by: Isaac <no-reply@databricks.com>
…ting on main) Co-authored-by: Isaac <no-reply@databricks.com>
Co-authored-by: Isaac <no-reply@databricks.com>
Co-authored-by: Isaac <no-reply@databricks.com>
…; record volume_id in empty-resource-id The volume's created_at/updated_at made rename_interrupted flaky ([UNIX_TIME_MILLIS] vs [UNIX_TIME_MILLIS][0] depending on timing). Co-authored-by: Isaac <no-reply@databricks.com>
Co-authored-by: Jan N Rose <janniklas.rose@gmail.com>
The recreate deletes the volume at its recorded path after the schema's rename moved it, so the moved volume is left behind. Co-authored-by: Isaac <no-reply@databricks.com>
denik
force-pushed
the
denik/models-rename
branch
from
November 10, 2026 14:14
36755c1 to
13aeadd
Compare
A parent renamed in the same plan moves the resource before its recreate runs. The recreate now deletes it at the ID the rename gives it, after the rename. The parent comes from the state's depends_on, since the config may no longer reference it. Co-authored-by: Isaac <no-reply@databricks.com>
Co-authored-by: Isaac <no-reply@databricks.com>
denik
force-pushed
the
denik/models-rename
branch
from
November 10, 2026 15:28
13aeadd to
a44c3dc
Compare
The plan's state_only re-keys, found IDs and added depends_on must survive serialization. Co-authored-by: Isaac <no-reply@databricks.com>
A registered model's name is now an updatable ID field: a rename sends new_name and keeps the model's versions, instead of deleting and recreating it. A model inside a renamed catalog or schema is re-keyed without a request, like volumes; recreating it failed with RESOURCE_ALREADY_EXISTS. Co-authored-by: Isaac <no-reply@databricks.com>
Co-authored-by: Isaac <no-reply@databricks.com>
After an interrupted parent rename the plan reads the model at its stale ID and has no remote state; sync aliases against the model at its new path instead. Co-authored-by: Isaac <no-reply@databricks.com>
…s with their schema A model missing from its recorded ID is planned as a create again, so DoUpdateWithID always has the plan's remote state. A forced schema or catalog delete in the test server now also deletes the registered models in it, like UC. Co-authored-by: Isaac <no-reply@databricks.com>
Co-authored-by: Isaac <no-reply@databricks.com>
Co-authored-by: Isaac <no-reply@databricks.com>
…; test rename with a comment change Co-authored-by: Isaac <no-reply@databricks.com>
A config read back from a saved plan force-sends name, so a rename deployed with --plan sent "name": "" next to new_name, which UC rejects. Co-authored-by: Isaac <no-reply@databricks.com>
Co-authored-by: Isaac <no-reply@databricks.com>
denik
force-pushed
the
denik/models-rename
branch
from
November 10, 2026 16:06
a44c3dc to
43d07f6
Compare
idelafuente-oc
pushed a commit
to idelafuente-oc/cli
that referenced
this pull request
Nov 11, 2026
…names (databricks#7408) ## Changes Make the test server behave like Unity Catalog where the UC rename PRs depend on it: - A renamed catalog moves its schemas, volumes, registered models and secrets. A renamed schema moves its volumes, registered models and secrets. - A forced catalog or schema delete also deletes what is in it. - Creating a catalog under a taken name fails with `CATALOG_ALREADY_EXISTS`. - A renamed schema or volume gets a lowercased name. A renamed registered model updates its `full_name` and keeps its versions, and setting `name` and `new_name` in one update is rejected. The `schema-catalog-change` and `registered-model-catalog-change` migrate scripts created a catalog that their captured replay creates again; the fake now rejects that, so the scripts no longer create it. ## Tests New `acceptance/cmd/workspace/uc-renames` (Cloud = true) checks each behavior with plain CLI calls. It passes locally and on AWS, GCP and Azure. Next: databricks#7343 (schema rename) stacks on this PR, followed by databricks#7379 and databricks#7381. This pull request and its description were written by Isaac. --------- Co-authored-by: Isaac <no-reply@databricks.com>
idelafuente-oc
pushed a commit
to idelafuente-oc/cli
that referenced
this pull request
Nov 14, 2026
… or schema (databricks#7343) ## Changes Changing a schema's `name` now renames it in place (`PATCH new_name`) instead of deleting and recreating it, so the schema and its contents survive. UC moves the children of a renamed catalog or schema along with it. A volume in a renamed schema, or a schema in a renamed catalog (`schema_name`/`catalog_name: ${resources.<parent>.name}`), used to be planned as a recreate, whose create then failed with `RESOURCE_ALREADY_EXISTS` on every deploy. The planner now re-keys such children (`update_id`, reason `parent_renamed`). Catalog rename was already supported and had the same latent bug for child schemas. - A child whose only change is following its parent is `state_only`: the framework saves it under the ID the adapter's new optional `IDFromState` gives it, with no request. It is left out of the text plan and counted as unchanged. A child with its own changes too goes through `DoUpdateWithID` at its new path. - A child recreated in the same deploy as its parent's rename is deleted where the rename moved it: the plan points the delete at the re-keyed ID (the entry's `id`) and orders it after the parent, taken from the state's `depends_on` because the config may now reference another one. Before, the delete went to the old path and the moved volume was left behind. - Registered models and UC secrets cannot follow a rename until databricks#7379 and databricks#7381, so in a renamed schema they are recreated where the rename moved them, as on main. A resource can also be missing from its recorded ID with no rename in the plan: a deploy stopped after renaming its parent, or it was renamed out of band. The plan then reads it at the `IDFromState` ID and takes what it finds only if its internal ID matches the recorded one, so a different object under that name is never adopted; the plan entry's `id` (declared, but never set until now) records where it is, and apply addresses it there. If that read fails with anything but not-found, the plan warns and goes on as before. - The new `internal_id_field` lifecycle setting names the internal ID (`schema_id`, `volume_id`). The framework records it from the remote state on every save, and the plan skips it (reason `internal_id`). - A state recorded without it (by an older CLI, or by `bind`) gets it from a state-only update. Like a skip, it needs no confirmation in `bind`, and references to it read the remote state. A migration from terraform records it directly: `schema_id` from the terraform state, `volume_id` from the remote (read by `bundle deployment migrate`, taken from the plan check when migrating on deploy). - With the new state fields, `apitypes.generated.yml` maps schemas and volumes to `SchemaInfo`/`VolumeInfo`, as registered models already map to their Info type. The test server now looks a volume's full name up case-insensitively and rejects a taken schema name, like UC. `state_present` and the synced-table terraform fixture no longer create a schema twice. The earlier test-server work is in databricks#7408 (merged). Known gaps, all as on main, recorded with `Badness`: destroy does not look for a resource renamed out of band (`schemas/destroy_after_rename_out_of_band`); the plan takes whatever is at the recorded name without checking the internal ID (`schemas/foreign_at_recorded_name`); a catalog renamed out of band is not found (`catalogs/rename_out_of_band`, fixed by databricks#7418). A parent named by a literal rather than a reference is not followed, nor is one interpolated into another field (a quality monitor's `table_name`). With databricks#7345 and databricks#7361 (merged), a renamed schema keeps its grants and permissions in sync. Registered models and UC secrets follow parent renames in databricks#7379 and databricks#7381. ## Tests Acceptance tests marked `Cloud = true` pass locally and on AWS, GCP and Azure. - Renames, `Cloud = true`: `schemas/rename`, `schemas/rename_with_volume`, `catalogs/rename` (catalog → schema → volume), `volumes/rename_catalog_and_schema` (two parent renames at once), `schemas/rename_with_volume_ref` (a job reading the volume's `full_name`, `id`, `volume_id`), `schemas/rename_then_change` (update, rename and recreate reach the new path, also after a second rename and a rename back). - Rename together with another change (one PATCH): `schemas/rename_with_comment`, `catalogs/rename_with_comment`. Case-only renames: `schemas/rename_case_only`, `volumes/rename_case_only` (UC keeps the lowercase name; a re-key under a renamed schema keeps the config's casing, which UC reads case-insensitively). - Recreate during a parent rename: `volumes/move_from_renamed_schema` (`Cloud = true`), `volumes/recreate_in_renamed_schema` (local only: an external volume needs an external location), `schemas/rename_with_model_and_secret` (`Cloud = true`). - Out of band: `volumes/rename_out_of_band`, `schemas/rename_out_of_band` (found by id, no requests), `catalogs/rename_out_of_band`, `catalogs/rename_all_out_of_band`, `volumes/recreate_out_of_band`, `schemas/recreate_out_of_band` (a different id: created, not adopted), `schemas/rename_out_of_band_old_state` (a state without the ids records them, then follows the rename), `volumes/old_state_ref` (a job reading the volume's fields stays unchanged while the ids are recorded). `deployment/bind/volume` now binds without `--auto-approve`. - Interrupted deploys (local only, kill hook): `schemas/rename_interrupted`, `schemas/recreate_interrupted`. A failed lookup read: `schemas/rename_probe_forbidden` (local only, injected 403). - The key rename tests also deploy from a saved plan (`EnvMatrix.READPLAN`). - Migration: `migrate/command/{basic,grants}` record the ids; the invariant `migrate` and `auto-migrate` targets show no drift. `grants/schemas/change_name` and `migrate/auto/schema-rename` show the in-place rename; `lifecycle/prevent-destroy` switched its recreate trigger from `name` to `catalog_name`. - Unit tests for `addParentMoves`, `followsParentRename`, `Changes.HasOnlySingleReason` and the skipped internal-ID changes. This pull request and its description were written by Isaac. --------- Co-authored-by: Isaac <no-reply@databricks.com> Co-authored-by: Jan N Rose <janniklas.rose@gmail.com>
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Changes
Changing a registered model's
namenow renames it in place (PATCH new_name) instead of deleting and recreating it, so the model keeps its versions. The recreate only matched Terraform, which is gone.A model in a renamed catalog or schema (
schema_name: ${resources.schemas.<s>.name}) now follows its parent the way volumes do in #7343: withIDFromState, the framework re-keys it without a request (state_only). It used to be planned as a recreate whose create failed withRESOURCE_ALREADY_EXISTS(verified on AWS), and every later deploy tried to create it again.UC rejects an update that sets both
nameandnew_name, so a rename leavesnameout (#7408 makes the test server reject it too).Stacked on #7343. UC secrets get the same fix in #7381.
Tests
registered_models/rename(Cloud = true): rename the model, then its schema, then its catalog, then the model again together with its comment (one PATCH).created_atstays the same and the final plan is a no-op.registered_models/basic(Cloud = true, AWS) andmigrate/auto/registered-model-rename: a name change now renames.This pull request and its description were written by Isaac.