Details

    • Type: Bug
    • Status: Closed - Won't Fix
    • Resolution: Cannot Reproduce
    • Affects Version/s: BIGDATA_RELEASE_1_4_0
    • Fix Version/s: None
    • Component/s: B+Tree
    • Labels:
      None

      Description

      One HAJournalServer appears to have hit a condition where 1 cpu core is at 100% in what appears to be a tight loop. I suspect an edge case in the code where hasPrior() is spinning.

      ?qtp1026925903-50" prio=10 tid=0x00007fe3bc95a800 nid=0x3e81 runnable [0x00007fe2a1fa4000]
         java.lang.Thread.State: RUNNABLE
      	at java.util.WeakHashMap.put(WeakHashMap.java:521)
      	at com.bigdata.btree.Leaf.addLeafListener(Leaf.java:2255)
      	at com.bigdata.btree.AbstractBTreeTupleCursor$MutableCursorPosition.<init>(AbstractBTreeTupleCursor.java:2054)
      	at com.bigdata.btree.AbstractBTreeTupleCursor$MutableBTreeTupleCursor.newPosition(AbstractBTreeTupleCursor.java:2238)
      	at com.bigdata.btree.AbstractBTreeTupleCursor$MutableBTreeTupleCursor.newPosition(AbstractBTreeTupleCursor.java:2224)
      	at com.bigdata.btree.AbstractBTreeTupleCursor.lastPosition(AbstractBTreeTupleCursor.java:615)
      	at com.bigdata.btree.AbstractBTreeTupleCursor.hasPrior(AbstractBTreeTupleCursor.java:1019)
      	at com.bigdata.btree.filter.Reverserator.hasNext(Reverserator.java:39)
      	at com.bigdata.btree.ResultSet.<init>(ResultSet.java:1060)
      	at com.bigdata.btree.ChunkedLocalRangeIterator.getResultSet(ChunkedLocalRangeIterator.java:141)
      	at com.bigdata.btree.UnisolatedReadWriteIndex$ChunkedIterator.getResultSet(UnisolatedReadWriteIndex.java:623)
      	at com.bigdata.btree.AbstractChunkedTupleIterator.rangeQuery(AbstractChunkedTupleIterator.java:305)
      	at com.bigdata.btree.AbstractChunkedTupleIterator.hasNext(AbstractChunkedTupleIterator.java:427)
      	at com.bigdata.journal.AbstractCommitTimeIndex.findByCommitCounter(AbstractCommitTimeIndex.java:450)
      	at com.bigdata.journal.jini.ha.HALogNexus.getHALogFileBytesSinceCommitCounter(HALogNexus.java:922)
      	at com.bigdata.journal.jini.ha.SnapshotManager.isReadyToSnapshot(SnapshotManager.java:1319)
      

        Activity

        Hide
        bryanthompson bryanthompson added a comment -

        There were 835 HALog files in the file system. The default BTree branching factor was overridden on the command line to 256 (rather than the default of 32). We might be able to do an exhaustive test of conditions to discover this spin.

        Show
        bryanthompson bryanthompson added a comment - There were 835 HALog files in the file system. The default BTree branching factor was overridden on the command line to 256 (rather than the default of 32). We might be able to do an exhaustive test of conditions to discover this spin.
        Hide
        bryanthompson bryanthompson added a comment -

        Attachment TestReverserator.java has been added with description: reverserator stress test

        Show
        bryanthompson bryanthompson added a comment - Attachment TestReverserator.java has been added with description: reverserator stress test
        Hide
        bryanthompson bryanthompson added a comment -

        We have written a stress test (attached to this ticket), but we have not been able to replicate this problem.

        Show
        bryanthompson bryanthompson added a comment - We have written a stress test (attached to this ticket), but we have not been able to replicate this problem.

          People

          • Assignee:
            thompsonbry thompsonbry
            Reporter:
            bryanthompson bryanthompson
          • Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved: