mirror of
https://github.com/neondatabase/neon.git
synced 2026-06-04 22:10:39 +00:00
pageserver: add method to dump keyspace in mgmt api client (#6145)
Part of getpage@lsn benchmark epic: https://github.com/neondatabase/neon/issues/5771
This commit is contained in:
committed by
GitHub
parent
8619e6295a
commit
47873470db
1
Cargo.lock
generated
1
Cargo.lock
generated
@@ -3167,6 +3167,7 @@ dependencies = [
|
||||
"enum-map",
|
||||
"hex",
|
||||
"postgres_ffi",
|
||||
"rand 0.8.5",
|
||||
"serde",
|
||||
"serde_json",
|
||||
"serde_with",
|
||||
|
||||
@@ -24,3 +24,4 @@ workspace_hack.workspace = true
|
||||
|
||||
[dev-dependencies]
|
||||
bincode.workspace = true
|
||||
rand.workspace = true
|
||||
|
||||
@@ -144,3 +144,37 @@ impl Key {
|
||||
pub fn is_rel_block_key(key: &Key) -> bool {
|
||||
key.field1 == 0x00 && key.field4 != 0
|
||||
}
|
||||
|
||||
impl std::str::FromStr for Key {
|
||||
type Err = anyhow::Error;
|
||||
|
||||
fn from_str(s: &str) -> std::result::Result<Self, Self::Err> {
|
||||
Self::from_hex(s)
|
||||
}
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use std::str::FromStr;
|
||||
|
||||
use crate::key::Key;
|
||||
|
||||
use rand::Rng;
|
||||
use rand::SeedableRng;
|
||||
|
||||
#[test]
|
||||
fn display_fromstr_bijection() {
|
||||
let mut rng = rand::rngs::StdRng::seed_from_u64(42);
|
||||
|
||||
let key = Key {
|
||||
field1: rng.gen(),
|
||||
field2: rng.gen(),
|
||||
field3: rng.gen(),
|
||||
field4: rng.gen(),
|
||||
field5: rng.gen(),
|
||||
field6: rng.gen(),
|
||||
};
|
||||
|
||||
assert_eq!(key, Key::from_str(&format!("{key}")).unwrap());
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1,11 +1,12 @@
|
||||
use crate::repository::{key_range_size, singleton_range, Key};
|
||||
use postgres_ffi::BLCKSZ;
|
||||
use std::ops::Range;
|
||||
|
||||
use crate::key::Key;
|
||||
|
||||
///
|
||||
/// Represents a set of Keys, in a compact form.
|
||||
///
|
||||
#[derive(Clone, Debug, Default)]
|
||||
#[derive(Clone, Debug, Default, PartialEq, Eq)]
|
||||
pub struct KeySpace {
|
||||
/// Contiguous ranges of keys that belong to the key space. In key order,
|
||||
/// and with no overlap.
|
||||
@@ -186,6 +187,33 @@ impl KeySpaceRandomAccum {
|
||||
}
|
||||
}
|
||||
|
||||
pub fn key_range_size(key_range: &Range<Key>) -> u32 {
|
||||
let start = key_range.start;
|
||||
let end = key_range.end;
|
||||
|
||||
if end.field1 != start.field1
|
||||
|| end.field2 != start.field2
|
||||
|| end.field3 != start.field3
|
||||
|| end.field4 != start.field4
|
||||
{
|
||||
return u32::MAX;
|
||||
}
|
||||
|
||||
let start = (start.field5 as u64) << 32 | start.field6 as u64;
|
||||
let end = (end.field5 as u64) << 32 | end.field6 as u64;
|
||||
|
||||
let diff = end - start;
|
||||
if diff > u32::MAX as u64 {
|
||||
u32::MAX
|
||||
} else {
|
||||
diff as u32
|
||||
}
|
||||
}
|
||||
|
||||
pub fn singleton_range(key: Key) -> Range<Key> {
|
||||
key..key.next()
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::*;
|
||||
@@ -5,6 +5,7 @@ use const_format::formatcp;
|
||||
/// Public API types
|
||||
pub mod control_api;
|
||||
pub mod key;
|
||||
pub mod keyspace;
|
||||
pub mod models;
|
||||
pub mod reltag;
|
||||
pub mod shard;
|
||||
|
||||
@@ -1,3 +1,5 @@
|
||||
pub mod partitioning;
|
||||
|
||||
use std::{
|
||||
collections::HashMap,
|
||||
num::{NonZeroU64, NonZeroUsize},
|
||||
|
||||
151
libs/pageserver_api/src/models/partitioning.rs
Normal file
151
libs/pageserver_api/src/models/partitioning.rs
Normal file
@@ -0,0 +1,151 @@
|
||||
use utils::lsn::Lsn;
|
||||
|
||||
#[derive(Debug, PartialEq, Eq)]
|
||||
pub struct Partitioning {
|
||||
pub keys: crate::keyspace::KeySpace,
|
||||
|
||||
pub at_lsn: Lsn,
|
||||
}
|
||||
|
||||
impl serde::Serialize for Partitioning {
|
||||
fn serialize<S>(&self, serializer: S) -> std::result::Result<S::Ok, S::Error>
|
||||
where
|
||||
S: serde::Serializer,
|
||||
{
|
||||
pub struct KeySpace<'a>(&'a crate::keyspace::KeySpace);
|
||||
|
||||
impl<'a> serde::Serialize for KeySpace<'a> {
|
||||
fn serialize<S>(&self, serializer: S) -> std::result::Result<S::Ok, S::Error>
|
||||
where
|
||||
S: serde::Serializer,
|
||||
{
|
||||
use serde::ser::SerializeSeq;
|
||||
let mut seq = serializer.serialize_seq(Some(self.0.ranges.len()))?;
|
||||
for kr in &self.0.ranges {
|
||||
seq.serialize_element(&KeyRange(kr))?;
|
||||
}
|
||||
seq.end()
|
||||
}
|
||||
}
|
||||
|
||||
use serde::ser::SerializeMap;
|
||||
let mut map = serializer.serialize_map(Some(2))?;
|
||||
map.serialize_key("keys")?;
|
||||
map.serialize_value(&KeySpace(&self.keys))?;
|
||||
map.serialize_key("at_lsn")?;
|
||||
map.serialize_value(&WithDisplay(&self.at_lsn))?;
|
||||
map.end()
|
||||
}
|
||||
}
|
||||
|
||||
pub struct WithDisplay<'a, T>(&'a T);
|
||||
|
||||
impl<'a, T: std::fmt::Display> serde::Serialize for WithDisplay<'a, T> {
|
||||
fn serialize<S>(&self, serializer: S) -> std::result::Result<S::Ok, S::Error>
|
||||
where
|
||||
S: serde::Serializer,
|
||||
{
|
||||
serializer.collect_str(&self.0)
|
||||
}
|
||||
}
|
||||
|
||||
pub struct KeyRange<'a>(&'a std::ops::Range<crate::key::Key>);
|
||||
|
||||
impl<'a> serde::Serialize for KeyRange<'a> {
|
||||
fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
|
||||
where
|
||||
S: serde::Serializer,
|
||||
{
|
||||
use serde::ser::SerializeTuple;
|
||||
let mut t = serializer.serialize_tuple(2)?;
|
||||
t.serialize_element(&WithDisplay(&self.0.start))?;
|
||||
t.serialize_element(&WithDisplay(&self.0.end))?;
|
||||
t.end()
|
||||
}
|
||||
}
|
||||
|
||||
impl<'a> serde::Deserialize<'a> for Partitioning {
|
||||
fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
|
||||
where
|
||||
D: serde::Deserializer<'a>,
|
||||
{
|
||||
pub struct KeySpace(crate::keyspace::KeySpace);
|
||||
|
||||
impl<'de> serde::Deserialize<'de> for KeySpace {
|
||||
fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
|
||||
where
|
||||
D: serde::Deserializer<'de>,
|
||||
{
|
||||
#[serde_with::serde_as]
|
||||
#[derive(serde::Deserialize)]
|
||||
#[serde(transparent)]
|
||||
struct Key(#[serde_as(as = "serde_with::DisplayFromStr")] crate::key::Key);
|
||||
|
||||
#[serde_with::serde_as]
|
||||
#[derive(serde::Deserialize)]
|
||||
struct Range(Key, Key);
|
||||
|
||||
let ranges: Vec<Range> = serde::Deserialize::deserialize(deserializer)?;
|
||||
Ok(Self(crate::keyspace::KeySpace {
|
||||
ranges: ranges
|
||||
.into_iter()
|
||||
.map(|Range(start, end)| (start.0..end.0))
|
||||
.collect(),
|
||||
}))
|
||||
}
|
||||
}
|
||||
|
||||
#[serde_with::serde_as]
|
||||
#[derive(serde::Deserialize)]
|
||||
struct De {
|
||||
keys: KeySpace,
|
||||
#[serde_as(as = "serde_with::DisplayFromStr")]
|
||||
at_lsn: Lsn,
|
||||
}
|
||||
|
||||
let de: De = serde::Deserialize::deserialize(deserializer)?;
|
||||
Ok(Self {
|
||||
at_lsn: de.at_lsn,
|
||||
keys: de.keys.0,
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::*;
|
||||
|
||||
#[test]
|
||||
fn test_serialization_roundtrip() {
|
||||
let reference = r#"
|
||||
{
|
||||
"keys": [
|
||||
[
|
||||
"000000000000000000000000000000000000",
|
||||
"000000000000000000000000000000000001"
|
||||
],
|
||||
[
|
||||
"000000067F00000001000000000000000000",
|
||||
"000000067F00000001000000000000000002"
|
||||
],
|
||||
[
|
||||
"030000000000000000000000000000000000",
|
||||
"030000000000000000000000000000000003"
|
||||
]
|
||||
],
|
||||
"at_lsn": "0/2240160"
|
||||
}
|
||||
"#;
|
||||
|
||||
let de: Partitioning = serde_json::from_str(reference).unwrap();
|
||||
|
||||
let ser = serde_json::to_string(&de).unwrap();
|
||||
|
||||
let ser_de: serde_json::Value = serde_json::from_str(&ser).unwrap();
|
||||
|
||||
assert_eq!(
|
||||
ser_de,
|
||||
serde_json::from_str::<'_, serde_json::Value>(reference).unwrap()
|
||||
);
|
||||
}
|
||||
}
|
||||
@@ -92,6 +92,22 @@ impl Client {
|
||||
.map_err(Error::ReceiveBody)
|
||||
}
|
||||
|
||||
pub async fn keyspace(
|
||||
&self,
|
||||
tenant_id: TenantId,
|
||||
timeline_id: TimelineId,
|
||||
) -> Result<pageserver_api::models::partitioning::Partitioning> {
|
||||
let uri = format!(
|
||||
"{}/v1/tenant/{tenant_id}/timeline/{timeline_id}/keyspace",
|
||||
self.mgmt_api_endpoint
|
||||
);
|
||||
self.get(&uri)
|
||||
.await?
|
||||
.json()
|
||||
.await
|
||||
.map_err(Error::ReceiveBody)
|
||||
}
|
||||
|
||||
async fn get<U: IntoUrl>(&self, uri: U) -> Result<reqwest::Response> {
|
||||
self.request(Method::GET, uri, ()).await
|
||||
}
|
||||
|
||||
@@ -1,4 +1,2 @@
|
||||
pub mod routes;
|
||||
pub use routes::make_router;
|
||||
|
||||
pub use pageserver_api::models;
|
||||
|
||||
@@ -28,10 +28,6 @@ use utils::http::endpoint::request_span;
|
||||
use utils::http::json::json_request_or_empty_body;
|
||||
use utils::http::request::{get_request_param, must_get_query_param, parse_query_param};
|
||||
|
||||
use super::models::{
|
||||
StatusResponse, TenantConfigRequest, TenantCreateRequest, TenantCreateResponse, TenantInfo,
|
||||
TimelineCreateRequest, TimelineGcRequest, TimelineInfo,
|
||||
};
|
||||
use crate::context::{DownloadBehavior, RequestContext};
|
||||
use crate::deletion_queue::DeletionQueueClient;
|
||||
use crate::metrics::{StorageTimeOperation, STORAGE_TIME_GLOBAL};
|
||||
@@ -51,6 +47,10 @@ use crate::tenant::timeline::Timeline;
|
||||
use crate::tenant::{LogicalSizeCalculationCause, PageReconstructError, TenantSharedResources};
|
||||
use crate::{config::PageServerConf, tenant::mgr};
|
||||
use crate::{disk_usage_eviction_task, tenant};
|
||||
use pageserver_api::models::{
|
||||
StatusResponse, TenantConfigRequest, TenantCreateRequest, TenantCreateResponse, TenantInfo,
|
||||
TimelineCreateRequest, TimelineGcRequest, TimelineInfo,
|
||||
};
|
||||
use utils::{
|
||||
auth::SwappableJwtAuth,
|
||||
generation::Generation,
|
||||
@@ -66,7 +66,7 @@ use utils::{
|
||||
};
|
||||
|
||||
// Imports only used for testing APIs
|
||||
use super::models::ConfigureFailpointsRequest;
|
||||
use pageserver_api::models::ConfigureFailpointsRequest;
|
||||
|
||||
// For APIs that require an Active tenant, how long should we block waiting for that state?
|
||||
// This is not functionally necessary (clients will retry), but avoids generating a lot of
|
||||
@@ -1524,69 +1524,6 @@ async fn timeline_collect_keyspace(
|
||||
let timeline_id: TimelineId = parse_request_param(&request, "timeline_id")?;
|
||||
check_permission(&request, Some(tenant_shard_id.tenant_id))?;
|
||||
|
||||
struct Partitioning {
|
||||
keys: crate::keyspace::KeySpace,
|
||||
|
||||
at_lsn: Lsn,
|
||||
}
|
||||
|
||||
impl serde::Serialize for Partitioning {
|
||||
fn serialize<S>(&self, serializer: S) -> std::result::Result<S::Ok, S::Error>
|
||||
where
|
||||
S: serde::Serializer,
|
||||
{
|
||||
use serde::ser::SerializeMap;
|
||||
let mut map = serializer.serialize_map(Some(2))?;
|
||||
map.serialize_key("keys")?;
|
||||
map.serialize_value(&KeySpace(&self.keys))?;
|
||||
map.serialize_key("at_lsn")?;
|
||||
map.serialize_value(&WithDisplay(&self.at_lsn))?;
|
||||
map.end()
|
||||
}
|
||||
}
|
||||
|
||||
struct WithDisplay<'a, T>(&'a T);
|
||||
|
||||
impl<'a, T: std::fmt::Display> serde::Serialize for WithDisplay<'a, T> {
|
||||
fn serialize<S>(&self, serializer: S) -> std::result::Result<S::Ok, S::Error>
|
||||
where
|
||||
S: serde::Serializer,
|
||||
{
|
||||
serializer.collect_str(&self.0)
|
||||
}
|
||||
}
|
||||
|
||||
struct KeySpace<'a>(&'a crate::keyspace::KeySpace);
|
||||
|
||||
impl<'a> serde::Serialize for KeySpace<'a> {
|
||||
fn serialize<S>(&self, serializer: S) -> std::result::Result<S::Ok, S::Error>
|
||||
where
|
||||
S: serde::Serializer,
|
||||
{
|
||||
use serde::ser::SerializeSeq;
|
||||
let mut seq = serializer.serialize_seq(Some(self.0.ranges.len()))?;
|
||||
for kr in &self.0.ranges {
|
||||
seq.serialize_element(&KeyRange(kr))?;
|
||||
}
|
||||
seq.end()
|
||||
}
|
||||
}
|
||||
|
||||
struct KeyRange<'a>(&'a std::ops::Range<crate::repository::Key>);
|
||||
|
||||
impl<'a> serde::Serialize for KeyRange<'a> {
|
||||
fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
|
||||
where
|
||||
S: serde::Serializer,
|
||||
{
|
||||
use serde::ser::SerializeTuple;
|
||||
let mut t = serializer.serialize_tuple(2)?;
|
||||
t.serialize_element(&WithDisplay(&self.0.start))?;
|
||||
t.serialize_element(&WithDisplay(&self.0.end))?;
|
||||
t.end()
|
||||
}
|
||||
}
|
||||
|
||||
let at_lsn: Option<Lsn> = parse_query_param(&request, "at_lsn")?;
|
||||
|
||||
async {
|
||||
@@ -1598,7 +1535,9 @@ async fn timeline_collect_keyspace(
|
||||
.await
|
||||
.map_err(|e| ApiError::InternalServerError(e.into()))?;
|
||||
|
||||
json_response(StatusCode::OK, Partitioning { keys, at_lsn })
|
||||
let res = pageserver_api::models::partitioning::Partitioning { keys, at_lsn };
|
||||
|
||||
json_response(StatusCode::OK, res)
|
||||
}
|
||||
.instrument(info_span!("timeline_collect_keyspace", tenant_id = %tenant_shard_id.tenant_id, shard_id = %tenant_shard_id.shard_slug(), %timeline_id))
|
||||
.await
|
||||
|
||||
@@ -10,7 +10,7 @@ pub mod deletion_queue;
|
||||
pub mod disk_usage_eviction_task;
|
||||
pub mod http;
|
||||
pub mod import_datadir;
|
||||
pub mod keyspace;
|
||||
pub use pageserver_api::keyspace;
|
||||
pub mod metrics;
|
||||
pub mod page_cache;
|
||||
pub mod page_service;
|
||||
|
||||
@@ -2,38 +2,11 @@ use crate::walrecord::NeonWalRecord;
|
||||
use anyhow::Result;
|
||||
use bytes::Bytes;
|
||||
use serde::{Deserialize, Serialize};
|
||||
use std::ops::{AddAssign, Range};
|
||||
use std::ops::AddAssign;
|
||||
use std::time::Duration;
|
||||
|
||||
pub use pageserver_api::key::{Key, KEY_SIZE};
|
||||
|
||||
pub fn key_range_size(key_range: &Range<Key>) -> u32 {
|
||||
let start = key_range.start;
|
||||
let end = key_range.end;
|
||||
|
||||
if end.field1 != start.field1
|
||||
|| end.field2 != start.field2
|
||||
|| end.field3 != start.field3
|
||||
|| end.field4 != start.field4
|
||||
{
|
||||
return u32::MAX;
|
||||
}
|
||||
|
||||
let start = (start.field5 as u64) << 32 | start.field6 as u64;
|
||||
let end = (end.field5 as u64) << 32 | end.field6 as u64;
|
||||
|
||||
let diff = end - start;
|
||||
if diff > u32::MAX as u64 {
|
||||
u32::MAX
|
||||
} else {
|
||||
diff as u32
|
||||
}
|
||||
}
|
||||
|
||||
pub fn singleton_range(key: Key) -> Range<Key> {
|
||||
key..key.next()
|
||||
}
|
||||
|
||||
/// A 'value' stored for a one Key.
|
||||
#[derive(Debug, Clone, Serialize, Deserialize)]
|
||||
#[cfg_attr(test, derive(PartialEq))]
|
||||
|
||||
Reference in New Issue
Block a user