indexer_scheduler fix limiter usage

indexer/crawler.go

@@ -17,8 +17,10 @@ import (
 )
 
 type CrawlDispatcher struct {
+	// from Crawl()
 	ingest chan *models.ActorInfo
 
+	// from AddToCatchupQueue()
 	catchup chan *crawlWork
 
 	complete chan models.Uid
@@ -143,7 +145,9 @@ func (c *CrawlDispatcher) mainLoop() {
 			pdsID := crawlJob.act.PDS
 			limiter := c.repoFetcher.GetOrCreateLimiterLazy(pdsID)
 			now := time.Now()
-			wouldDelay := limiter.ReserveN(now, 1).DelayFrom(now)
+			res := limiter.ReserveN(now, 1)
+			wouldDelay := res.DelayFrom(now)
+			res.Cancel()
 			crawlJob.eligibleTime = now.Add(wouldDelay)
 			// put crawl job on heap sorted by eligible time
 			c.enheapJob(crawlJob)
@@ -318,6 +322,7 @@ func (c *CrawlDispatcher) enheapJob(crawlJob *crawlWork) {
 	defer c.repoSyncLock.Unlock()
 	pdsJobs, has := c.repoSyncPds[crawlJob.act.PDS]
 	if has {
+		// to honor PDS rate limits, put new job at end of per-pds list
 		if !pdsJobs.alreadyEnheaped {
 			heap.Push(c, crawlJob)
 			catchupPending.Set(float64(len(c.repoSyncHeap)))