compare

Author: wxiaoguang

.github/workflows/pull-db-tests.yml

@@ -160,12 +160,12 @@ jobs:
           ALLOW_EMPTY_PASSWORD: true
           MYSQL_DATABASE: testgitea
           # https://dev.mysql.com/doc/refman/8.0/en/optimizing-innodb-diskio.html
-          MYSQL_EXTRA_FLAGS: >
-            --disable-log-bin
-            --innodb-buffer-pool-size=1G
-            --innodb-doublewrite=off
-            --innodb-flush-log-at-trx-commit=0
-            --innodb-flush-method=nosync
+          #MYSQL_EXTRA_FLAGS: >
+          #  --disable-log-bin
+          #  --innodb-buffer-pool-size=1G
+          #  --innodb-doublewrite=off
+          #  --innodb-flush-log-at-trx-commit=0
+          #  --innodb-flush-method=nosync
         ports:
           - "3306:3306"
       elasticsearch:

modules/queue/workergroup.go

@@ -108,6 +108,8 @@ func (q *WorkerPoolQueue[T]) doWorkerHandle(batch []T) {
 			return // do not requeue items when flushing, since all items failed, requeue them will continue failing.
 		}
 		log.Error("Queue %q failed to handle batch of %d items, backoff for a few seconds", q.GetName(), len(batch))
+		// TODO: ideally it shouldn't "sleep" here (blocks the worker, then blocks flush).
+		// It could debounce the requeue operation, and try to requeue the items in the future.
 		select {
 		case <-q.ctxRun.Done():
 		case <-time.After(time.Duration(unhandledItemRequeueDuration.Load())):