Add a way to signal shutdown to snapshotting threads (#10744)

* Add a way to signal shutdown to snapshotting threads

* Pass Progress to fat_rlps() so we can abort from there too.

* Checking for abort in a single spot

* Remove nightly-only weak/strong counts

* fix warning

* Fix tests

* Add dummy impl to abort snapshots

* Add another dummy impl for TestSnapshotService

* Remove debugging code

* Return error instead of the odd Ok(())
Switch to AtomicU64
This commit is contained in:
David
2019-06-19 10:13:09 +02:00
committed by Andronik Ordian
parent be5db14160
commit d2120ded56
15 changed files with 167 additions and 78 deletions

View File

@@ -15,7 +15,6 @@
// along with Parity Ethereum. If not, see <http://www.gnu.org/licenses/>.
//! Ethcore client application.
#![warn(missing_docs)]
extern crate ansi_term;

View File

@@ -900,17 +900,27 @@ impl RunningClient {
// Create a weak reference to the client so that we can wait on shutdown
// until it is dropped
let weak_client = Arc::downgrade(&client);
// Shutdown and drop the ServiceClient
// Shutdown and drop the ClientService
client_service.shutdown();
trace!(target: "shutdown", "ClientService shut down");
drop(client_service);
trace!(target: "shutdown", "ClientService dropped");
// drop this stuff as soon as exit detected.
drop(rpc);
trace!(target: "shutdown", "RPC dropped");
drop(keep_alive);
trace!(target: "shutdown", "KeepAlive dropped");
// to make sure timer does not spawn requests while shutdown is in progress
informant.shutdown();
trace!(target: "shutdown", "Informant shut down");
// just Arc is dropping here, to allow other reference release in its default time
drop(informant);
trace!(target: "shutdown", "Informant dropped");
drop(client);
trace!(target: "shutdown", "Client dropped");
// This may help when debugging ref cycles. Requires nightly-only `#![feature(weak_counts)]`
// trace!(target: "shutdown", "Waiting for refs to Client to shutdown, strong_count={:?}, weak_count={:?}", weak_client.strong_count(), weak_client.weak_count());
trace!(target: "shutdown", "Waiting for refs to Client to shutdown");
wait_for_drop(weak_client);
}
}
@@ -944,24 +954,30 @@ fn print_running_environment(data_dir: &str, dirs: &Directories, db_dirs: &Datab
}
fn wait_for_drop<T>(w: Weak<T>) {
let sleep_duration = Duration::from_secs(1);
let warn_timeout = Duration::from_secs(60);
let max_timeout = Duration::from_secs(300);
const SLEEP_DURATION: Duration = Duration::from_secs(1);
const WARN_TIMEOUT: Duration = Duration::from_secs(60);
const MAX_TIMEOUT: Duration = Duration::from_secs(300);
let instant = Instant::now();
let mut warned = false;
while instant.elapsed() < max_timeout {
while instant.elapsed() < MAX_TIMEOUT {
if w.upgrade().is_none() {
return;
}
if !warned && instant.elapsed() > warn_timeout {
if !warned && instant.elapsed() > WARN_TIMEOUT {
warned = true;
warn!("Shutdown is taking longer than expected.");
}
thread::sleep(sleep_duration);
thread::sleep(SLEEP_DURATION);
// When debugging shutdown issues on a nightly build it can help to enable this with the
// `#![feature(weak_counts)]` added to lib.rs (TODO: enable when
// https://github.com/rust-lang/rust/issues/57977 is stable)
// trace!(target: "shutdown", "Waiting for client to drop, strong_count={:?}, weak_count={:?}", w.strong_count(), w.weak_count());
trace!(target: "shutdown", "Waiting for client to drop");
}
warn!("Shutdown timeout reached, exiting uncleanly.");

View File

@@ -261,7 +261,7 @@ impl SnapshotCommand {
let cur_size = p.size();
if cur_size != last_size {
last_size = cur_size;
let bytes = ::informant::format_bytes(p.size());
let bytes = ::informant::format_bytes(cur_size as usize);
info!("Snapshot: {} accounts {} blocks {}", p.accounts(), p.blocks(), bytes);
}