pageserver: move deletion failpoint inside backoff (#5814)

## Problem

When enabled, this failpoint would busy-spin in a loop that emits log
messages.

## Summary of changes

Move the failpoint inside a backoff::exponential block: it will still
spam the log, but at much lower rate.

---------

Co-authored-by: Joonas Koivunen <joonas@neon.tech>
This commit is contained in:
John Spray
2023-11-07 14:25:51 +00:00
committed by GitHub
parent 4cd47b7d4b
commit 1d68f52b57

View File

@@ -55,21 +55,24 @@ impl Deleter {
/// Wrap the remote `delete_objects` with a failpoint
async fn remote_delete(&self) -> Result<(), anyhow::Error> {
fail::fail_point!("deletion-queue-before-execute", |_| {
info!("Skipping execution, failpoint set");
metrics::DELETION_QUEUE
.remote_errors
.with_label_values(&["failpoint"])
.inc();
Err(anyhow::anyhow!("failpoint hit"))
});
// A backoff::retry is used here for two reasons:
// - To provide a backoff rather than busy-polling the API on errors
// - To absorb transient 429/503 conditions without hitting our error
// logging path for issues deleting objects.
backoff::retry(
|| async { self.remote_storage.delete_objects(&self.accumulator).await },
|| async {
fail::fail_point!("deletion-queue-before-execute", |_| {
info!("Skipping execution, failpoint set");
metrics::DELETION_QUEUE
.remote_errors
.with_label_values(&["failpoint"])
.inc();
Err(anyhow::anyhow!("failpoint: deletion-queue-before-execute"))
});
self.remote_storage.delete_objects(&self.accumulator).await
},
|_| false,
3,
10,