[compute_ctl] Do not create availability checker data on each start (#4019)

Initially, idea was to ensure that when we come and check data
availability, special service table already contains one row. So if we
loose it for some reason, we will error out.

Yet, to do availability check we anyway start compute first! So it
doesn't really add some value, but we affect each compute start as we
update at least one row in the database. Also this writes some WAL, so
if timeline is close to `neon.max_cluster_size` it could prevent compute
from starting up.

That said, do CREATE TABLE IF NOT EXISTS + UPSERT right in the
`/check_writability` handler.
This commit is contained in:
Alexey Kondratov
2023-04-14 13:05:07 +02:00
committed by GitHub
parent 0c82ff3d98
commit 589cf1ed21
3 changed files with 31 additions and 30 deletions

View File

@@ -1,12 +1,28 @@
use anyhow::{anyhow, Result};
use postgres::Client;
use tokio_postgres::NoTls;
use tracing::{error, instrument};
use crate::compute::ComputeNode;
/// Update timestamp in a row in a special service table to check
/// that we can actually write some data in this particular timeline.
/// Create table if it's missing.
#[instrument(skip_all)]
pub fn create_writability_check_data(client: &mut Client) -> Result<()> {
pub async fn check_writability(compute: &ComputeNode) -> Result<()> {
// Connect to the database.
let (client, connection) = tokio_postgres::connect(compute.connstr.as_str(), NoTls).await?;
if client.is_closed() {
return Err(anyhow!("connection to postgres closed"));
}
// The connection object performs the actual communication with the database,
// so spawn it off to run on its own.
tokio::spawn(async move {
if let Err(e) = connection.await {
error!("connection error: {}", e);
}
});
let query = "
CREATE TABLE IF NOT EXISTS health_check (
id serial primary key,
@@ -15,31 +31,15 @@ pub fn create_writability_check_data(client: &mut Client) -> Result<()> {
INSERT INTO health_check VALUES (1, now())
ON CONFLICT (id) DO UPDATE
SET updated_at = now();";
let result = client.simple_query(query)?;
if result.len() < 2 {
return Err(anyhow::format_err!("executed {} queries", result.len()));
}
Ok(())
}
#[instrument(skip_all)]
pub async fn check_writability(compute: &ComputeNode) -> Result<()> {
let (client, connection) = tokio_postgres::connect(compute.connstr.as_str(), NoTls).await?;
if client.is_closed() {
return Err(anyhow!("connection to postgres closed"));
}
tokio::spawn(async move {
if let Err(e) = connection.await {
error!("connection error: {}", e);
}
});
let result = client
.simple_query("UPDATE health_check SET updated_at = now() WHERE id = 1;")
.await?;
if result.len() != 1 {
return Err(anyhow!("statement can't be executed"));
let result = client.simple_query(query).await?;
if result.len() != 2 {
return Err(anyhow::format_err!(
"expected 2 query results, but got {}",
result.len()
));
}
Ok(())
}

View File

@@ -32,7 +32,6 @@ use utils::lsn::Lsn;
use compute_api::responses::{ComputeMetrics, ComputeStatus};
use compute_api::spec::ComputeSpec;
use crate::checker::create_writability_check_data;
use crate::config;
use crate::pg_helpers::*;
use crate::spec::*;
@@ -342,7 +341,6 @@ impl ComputeNode {
handle_databases(spec, &mut client)?;
handle_role_deletions(spec, self.connstr.as_str(), &mut client)?;
handle_grants(spec, self.connstr.as_str(), &mut client)?;
create_writability_check_data(&mut client)?;
handle_extensions(spec, &mut client)?;
// 'Close' connection

View File

@@ -85,7 +85,10 @@ async fn routes(req: Request<Body>, compute: &Arc<ComputeNode>) -> Response<Body
let res = crate::checker::check_writability(compute).await;
match res {
Ok(_) => Response::new(Body::from("true")),
Err(e) => Response::new(Body::from(e.to_string())),
Err(e) => {
error!("check_writability failed: {}", e);
Response::new(Body::from(e.to_string()))
}
}
}