From fa2e5498a528e267eaca111187b80b2353b44bf4 Mon Sep 17 00:00:00 2001 From: John Spray Date: Thu, 13 Feb 2025 22:06:29 +0100 Subject: [PATCH] storcon: don't enqueue reconciles on failed startup compute hook --- storage_controller/src/service.rs | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/storage_controller/src/service.rs b/storage_controller/src/service.rs index 4028cd7023..e1f2d60c25 100644 --- a/storage_controller/src/service.rs +++ b/storage_controller/src/service.rs @@ -5034,12 +5034,13 @@ impl Service { // If we failed any compute notifications, make a note to retry later. if !failed_notifications.is_empty() { - let mut locked = self.inner.write().unwrap(); - for failed in failed_notifications { - if let Some(shard) = locked.tenants.get_mut(&failed) { - shard.pending_compute_notification = true; - } - } + tracing::warn!("Failed to notify compute of {} shards, not enqueueing for retry to avoid blocking other work. Some computes might miss updates.", failed_notifications.len()); + // let mut locked = self.inner.write().unwrap(); + // for failed in failed_notifications { + // if let Some(shard) = locked.tenants.get_mut(&failed) { + // shard.pending_compute_notification = true; + // } + // } } Ok((response, waiters))