chore(compute_ctl): add error context to apply_spec (#7374)

Make it faster to identify which part of apply spec goes wrong by adding
an error context.

Signed-off-by: Alex Chi Z <chi@neon.tech>
This commit is contained in:
Alex Chi Z
2024-04-17 09:11:04 +03:00
committed by GitHub
parent 9e567d9814
commit cb4b40f9c1
2 changed files with 60 additions and 34 deletions

View File

@@ -818,9 +818,15 @@ impl ComputeNode {
Client::connect(zenith_admin_connstr.as_str(), NoTls)
.context("broken cloud_admin credential: tried connecting with cloud_admin but could not authenticate, and zenith_admin does not work either")?;
// Disable forwarding so that users don't get a cloud_admin role
client.simple_query("SET neon.forward_ddl = false")?;
client.simple_query("CREATE USER cloud_admin WITH SUPERUSER")?;
client.simple_query("GRANT zenith_admin TO cloud_admin")?;
let mut func = || {
client.simple_query("SET neon.forward_ddl = false")?;
client.simple_query("CREATE USER cloud_admin WITH SUPERUSER")?;
client.simple_query("GRANT zenith_admin TO cloud_admin")?;
Ok::<_, anyhow::Error>(())
};
func().context("apply_config setup cloud_admin")?;
drop(client);
// reconnect with connstring with expected name
@@ -832,24 +838,29 @@ impl ComputeNode {
};
// Disable DDL forwarding because control plane already knows about these roles/databases.
client.simple_query("SET neon.forward_ddl = false")?;
client
.simple_query("SET neon.forward_ddl = false")
.context("apply_config SET neon.forward_ddl = false")?;
// Proceed with post-startup configuration. Note, that order of operations is important.
let spec = &compute_state.pspec.as_ref().expect("spec must be set").spec;
create_neon_superuser(spec, &mut client)?;
cleanup_instance(&mut client)?;
handle_roles(spec, &mut client)?;
handle_databases(spec, &mut client)?;
handle_role_deletions(spec, connstr.as_str(), &mut client)?;
create_neon_superuser(spec, &mut client).context("apply_config create_neon_superuser")?;
cleanup_instance(&mut client).context("apply_config cleanup_instance")?;
handle_roles(spec, &mut client).context("apply_config handle_roles")?;
handle_databases(spec, &mut client).context("apply_config handle_databases")?;
handle_role_deletions(spec, connstr.as_str(), &mut client)
.context("apply_config handle_role_deletions")?;
handle_grants(
spec,
&mut client,
connstr.as_str(),
self.has_feature(ComputeFeature::AnonExtension),
)?;
handle_extensions(spec, &mut client)?;
handle_extension_neon(&mut client)?;
create_availability_check_data(&mut client)?;
)
.context("apply_config handle_grants")?;
handle_extensions(spec, &mut client).context("apply_config handle_extensions")?;
handle_extension_neon(&mut client).context("apply_config handle_extension_neon")?;
create_availability_check_data(&mut client)
.context("apply_config create_availability_check_data")?;
// 'Close' connection
drop(client);
@@ -857,7 +868,7 @@ impl ComputeNode {
// Run migrations separately to not hold up cold starts
thread::spawn(move || {
let mut client = Client::connect(connstr.as_str(), NoTls)?;
handle_migrations(&mut client)
handle_migrations(&mut client).context("apply_config handle_migrations")
});
Ok(())
}

View File

@@ -2,7 +2,7 @@ use std::fs::File;
use std::path::Path;
use std::str::FromStr;
use anyhow::{anyhow, bail, Result};
use anyhow::{anyhow, bail, Context, Result};
use postgres::config::Config;
use postgres::{Client, NoTls};
use reqwest::StatusCode;
@@ -698,7 +698,8 @@ pub fn handle_grants(
// it is important to run this after all grants
if enable_anon_extension {
handle_extension_anon(spec, &db.owner, &mut db_client, false)?;
handle_extension_anon(spec, &db.owner, &mut db_client, false)
.context("handle_grants handle_extension_anon")?;
}
}
@@ -813,28 +814,36 @@ $$;"#,
// Add new migrations below.
];
let mut query = "CREATE SCHEMA IF NOT EXISTS neon_migration";
client.simple_query(query)?;
let mut func = || {
let query = "CREATE SCHEMA IF NOT EXISTS neon_migration";
client.simple_query(query)?;
query = "CREATE TABLE IF NOT EXISTS neon_migration.migration_id (key INT NOT NULL PRIMARY KEY, id bigint NOT NULL DEFAULT 0)";
client.simple_query(query)?;
let query = "CREATE TABLE IF NOT EXISTS neon_migration.migration_id (key INT NOT NULL PRIMARY KEY, id bigint NOT NULL DEFAULT 0)";
client.simple_query(query)?;
query = "INSERT INTO neon_migration.migration_id VALUES (0, 0) ON CONFLICT DO NOTHING";
client.simple_query(query)?;
let query = "INSERT INTO neon_migration.migration_id VALUES (0, 0) ON CONFLICT DO NOTHING";
client.simple_query(query)?;
query = "ALTER SCHEMA neon_migration OWNER TO cloud_admin";
client.simple_query(query)?;
let query = "ALTER SCHEMA neon_migration OWNER TO cloud_admin";
client.simple_query(query)?;
query = "REVOKE ALL ON SCHEMA neon_migration FROM PUBLIC";
client.simple_query(query)?;
let query = "REVOKE ALL ON SCHEMA neon_migration FROM PUBLIC";
client.simple_query(query)?;
Ok::<_, anyhow::Error>(())
};
func().context("handle_migrations prepare")?;
query = "SELECT id FROM neon_migration.migration_id";
let row = client.query_one(query, &[])?;
let query = "SELECT id FROM neon_migration.migration_id";
let row = client
.query_one(query, &[])
.context("handle_migrations get migration_id")?;
let mut current_migration: usize = row.get::<&str, i64>("id") as usize;
let starting_migration_id = current_migration;
query = "BEGIN";
client.simple_query(query)?;
let query = "BEGIN";
client
.simple_query(query)
.context("handle_migrations begin")?;
while current_migration < migrations.len() {
let migration = &migrations[current_migration];
@@ -842,7 +851,9 @@ $$;"#,
info!("Skip migration id={}", current_migration);
} else {
info!("Running migration:\n{}\n", migration);
client.simple_query(migration)?;
client.simple_query(migration).with_context(|| {
format!("handle_migrations current_migration={}", current_migration)
})?;
}
current_migration += 1;
}
@@ -850,10 +861,14 @@ $$;"#,
"UPDATE neon_migration.migration_id SET id={}",
migrations.len()
);
client.simple_query(&setval)?;
client
.simple_query(&setval)
.context("handle_migrations update id")?;
query = "COMMIT";
client.simple_query(query)?;
let query = "COMMIT";
client
.simple_query(query)
.context("handle_migrations commit")?;
info!(
"Ran {} migrations",