pageserver: location configuration API, attachment modes, secondary l…

…ocations (#5299)

## Problem

These changes are part of building seamless tenant migration, as
described in the RFC:
- #5029

## Summary of changes

- A new configuration type `LocationConf` supersedes `TenantConfOpt` for
storing a tenant's configuration in the pageserver repo dir. It contains
`TenantConfOpt`, as well as a new `mode` attribute that describes what
kind of location this is (secondary, attached, attachment mode etc). It
is written to a file called `config-v1` instead of `config` -- this
prepares us for neatly making any other profound changes to the format
of the file in future. Forward compat for existing pageserver code is
achieved by writing out both old and new style files. Backward compat is
achieved by checking for the old-style file if the new one isn't found.
- The `TenantMap` type changes, to hold `TenantSlot` instead of just
`Tenant`. The `Tenant` type continues to be used for attached tenants
only. Tenants in other states (such as secondaries) are represented by a
different variant of `TenantSlot`.
- Where `Tenant` & `Timeline` used to hold an Arc<Mutex<TenantConfOpt>>,
they now hold a reference to a AttachedTenantConf, which includes the
extra information from LocationConf. This enables them to know the
current attachment mode.
- The attachment mode is used as an advisory input to decide whether to
do compaction and GC (AttachedStale is meant to avoid doing uploads,
AttachedMulti is meant to avoid doing deletions).
- A new HTTP API is added at `PUT /tenants/<tenant_id>/location_config`
to drive new location configuration. This provides a superset of the
functionality of attach/detach/load/ignore:
  - Attaching a tenant is just configuring it in an attached state
  - Detaching a tenant is configuring it to a detached state
  - Loading a tenant is just the same as attaching it
- Ignoring a tenant is the same as configuring it into Secondary with
warm=false (i.e. retain the files on disk but do nothing else).

Caveats:
- AttachedMulti tenants don't do compaction in this PR, but they do in
the follow on #5397
- Concurrent updates to the `location_config` API are not handled
elegantly in this PR, a better mechanism is added in the follow on
#5367
- Secondary mode is just a placeholder in this PR: the code to upload
heatmaps and do downloads on secondary locations will be added in a
later PR (but that shouldn't change any external interfaces)

Closes: #5379

---------

Co-authored-by: Christian Schwarz <christian@neon.tech>

libs/pageserver_api/src/models.rs

@@ -10,6 +10,7 @@ use serde_with::{serde_as, DisplayFromStr};
 use strum_macros;
 use utils::{
     completion,
+    generation::Generation,
     history_buffer::HistoryBufferWithDropCounter,
     id::{NodeId, TenantId, TimelineId},
     lsn::Lsn,
@@ -218,6 +219,8 @@ impl std::ops::Deref for TenantCreateRequest {
     }
 }
 
+/// An alternative representation of `pageserver::tenant::TenantConf` with
+/// simpler types.
 #[derive(Serialize, Deserialize, Debug, Default)]
 pub struct TenantConfig {
     pub checkpoint_distance: Option<u64>,
@@ -243,6 +246,39 @@ pub struct TenantConfig {
     pub gc_feedback: Option<bool>,
 }
 
+/// A flattened analog of a `pagesever::tenant::LocationMode`, which
+/// lists out all possible states (and the virtual "Detached" state)
+/// in a flat form rather than using rust-style enums.
+#[derive(Serialize, Deserialize, Debug)]
+pub enum LocationConfigMode {
+    AttachedSingle,
+    AttachedMulti,
+    AttachedStale,
+    Secondary,
+    Detached,
+}
+
+#[derive(Serialize, Deserialize, Debug)]
+pub struct LocationConfigSecondary {
+    pub warm: bool,
+}
+
+/// An alternative representation of `pageserver::tenant::LocationConf`,
+/// for use in external-facing APIs.
+#[derive(Serialize, Deserialize, Debug)]
+pub struct LocationConfig {
+    pub mode: LocationConfigMode,
+    /// If attaching, in what generation?
+    #[serde(default)]
+    pub generation: Option<Generation>,
+    #[serde(default)]
+    pub secondary_conf: Option<LocationConfigSecondary>,
+
+    // If requesting mode `Secondary`, configuration for that.
+    // Custom storage configuration for the tenant, if any
+    pub tenant_conf: TenantConfig,
+}
+
 #[serde_as]
 #[derive(Serialize, Deserialize)]
 #[serde(transparent)]
@@ -253,6 +289,16 @@ pub struct StatusResponse {
     pub id: NodeId,
 }
 
+#[serde_as]
+#[derive(Serialize, Deserialize, Debug)]
+#[serde(deny_unknown_fields)]
+pub struct TenantLocationConfigRequest {
+    #[serde_as(as = "DisplayFromStr")]
+    pub tenant_id: TenantId,
+    #[serde(flatten)]
+    pub config: LocationConfig, // as we have a flattened field, we should reject all unknown fields in it
+}
+
 #[serde_as]
 #[derive(Serialize, Deserialize, Debug)]
 #[serde(deny_unknown_fields)]

pageserver/src/config.rs

@@ -37,8 +37,8 @@ use crate::tenant::{
     TIMELINES_SEGMENT_NAME,
 };
 use crate::{
-    IGNORED_TENANT_FILE_NAME, METADATA_FILE_NAME, TENANT_CONFIG_NAME, TIMELINE_DELETE_MARK_SUFFIX,
-    TIMELINE_UNINIT_MARK_SUFFIX,
+    IGNORED_TENANT_FILE_NAME, METADATA_FILE_NAME, TENANT_CONFIG_NAME, TENANT_LOCATION_CONFIG_NAME,
+    TIMELINE_DELETE_MARK_SUFFIX, TIMELINE_UNINIT_MARK_SUFFIX,
 };
 
 pub mod defaults {
@@ -631,10 +631,18 @@ impl PageServerConf {
 
     /// Points to a place in pageserver's local directory,
     /// where certain tenant's tenantconf file should be located.
+    ///
+    /// Legacy: superseded by tenant_location_config_path.  Eventually
+    /// remove this function.
     pub fn tenant_config_path(&self, tenant_id: &TenantId) -> Utf8PathBuf {
         self.tenant_path(tenant_id).join(TENANT_CONFIG_NAME)
     }
 
+    pub fn tenant_location_config_path(&self, tenant_id: &TenantId) -> Utf8PathBuf {
+        self.tenant_path(tenant_id)
+            .join(TENANT_LOCATION_CONFIG_NAME)
+    }
+
     pub fn timelines_path(&self, tenant_id: &TenantId) -> Utf8PathBuf {
         self.tenant_path(tenant_id).join(TIMELINES_SEGMENT_NAME)
     }

pageserver/src/http/routes.rs

@@ -10,7 +10,8 @@ use hyper::StatusCode;
 use hyper::{Body, Request, Response, Uri};
 use metrics::launch_timestamp::LaunchTimestamp;
 use pageserver_api::models::{
-    DownloadRemoteLayersTaskSpawnRequest, TenantAttachRequest, TenantLoadRequest,
+    DownloadRemoteLayersTaskSpawnRequest, LocationConfigMode, TenantAttachRequest,
+    TenantLoadRequest, TenantLocationConfigRequest,
 };
 use remote_storage::GenericRemoteStorage;
 use tenant_size_model::{SizeResult, StorageModel};
@@ -29,7 +30,7 @@ use crate::deletion_queue::DeletionQueueClient;
 use crate::metrics::{StorageTimeOperation, STORAGE_TIME_GLOBAL};
 use crate::pgdatadir_mapping::LsnForTimestamp;
 use crate::task_mgr::TaskKind;
-use crate::tenant::config::TenantConfOpt;
+use crate::tenant::config::{LocationConf, TenantConfOpt};
 use crate::tenant::mgr::{
     GetTenantError, SetNewTenantConfigError, TenantMapInsertError, TenantStateError,
 };
@@ -150,7 +151,10 @@ impl From<TenantMapInsertError> for ApiError {
             TenantMapInsertError::TenantAlreadyExists(id, state) => {
                 ApiError::Conflict(format!("tenant {id} already exists, state: {state:?}"))
             }
-            TenantMapInsertError::Closure(e) => ApiError::InternalServerError(e),
+            TenantMapInsertError::TenantExistsSecondary(id) => {
+                ApiError::Conflict(format!("tenant {id} already exists as secondary"))
+            }
+            TenantMapInsertError::Other(e) => ApiError::InternalServerError(e),
         }
     }
 }
@@ -1011,6 +1015,48 @@ async fn update_tenant_config_handler(
     json_response(StatusCode::OK, ())
 }
 
+async fn put_tenant_location_config_handler(
+    mut request: Request<Body>,
+    _cancel: CancellationToken,
+) -> Result<Response<Body>, ApiError> {
+    let request_data: TenantLocationConfigRequest = json_request(&mut request).await?;
+    let tenant_id = request_data.tenant_id;
+    check_permission(&request, Some(tenant_id))?;
+
+    let ctx = RequestContext::new(TaskKind::MgmtRequest, DownloadBehavior::Warn);
+    let state = get_state(&request);
+    let conf = state.conf;
+
+    // The `Detached` state is special, it doesn't upsert a tenant, it removes
+    // its local disk content and drops it from memory.
+    if let LocationConfigMode::Detached = request_data.config.mode {
+        mgr::detach_tenant(conf, tenant_id, true)
+            .instrument(info_span!("tenant_detach", %tenant_id))
+            .await?;
+        return json_response(StatusCode::OK, ());
+    }
+
+    let location_conf =
+        LocationConf::try_from(&request_data.config).map_err(ApiError::BadRequest)?;
+
+    mgr::upsert_location(
+        state.conf,
+        tenant_id,
+        location_conf,
+        state.broker_client.clone(),
+        state.remote_storage.clone(),
+        state.deletion_queue_client.clone(),
+        &ctx,
+    )
+    .await
+    // TODO: badrequest assumes the caller was asking for something unreasonable, but in
+    // principle we might have hit something like concurrent API calls to the same tenant,
+    // which is not a 400 but a 409.
+    .map_err(ApiError::BadRequest)?;
+
+    json_response(StatusCode::OK, ())
+}
+
 /// Testing helper to transition a tenant to [`crate::tenant::TenantState::Broken`].
 async fn handle_tenant_break(
     r: Request<Body>,
@@ -1464,6 +1510,9 @@ pub fn make_router(
         .get("/v1/tenant/:tenant_id/config", |r| {
             api_handler(r, get_tenant_config_handler)
         })
+        .put("/v1/tenant/:tenant_id/location_config", |r| {
+            api_handler(r, put_tenant_location_config_handler)
+        })
         .get("/v1/tenant/:tenant_id/timeline", |r| {
             api_handler(r, timeline_list_handler)
         })

pageserver/src/lib.rs

@@ -112,6 +112,10 @@ pub const METADATA_FILE_NAME: &str = "metadata";
 /// Full path: `tenants/<tenant_id>/config`.
 pub const TENANT_CONFIG_NAME: &str = "config";
 
+/// Per-tenant configuration file.
+/// Full path: `tenants/<tenant_id>/config`.
+pub const TENANT_LOCATION_CONFIG_NAME: &str = "config-v1";
+
 /// A suffix used for various temporary files. Any temporary files found in the
 /// data directory at pageserver startup can be automatically removed.
 pub const TEMP_FILE_SUFFIX: &str = "___temp";