accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Josh Elser (JIRA)" <j...@apache.org>
Subject [jira] [Created] (ACCUMULO-3877) TableOperationsIT failed in testCompactEmptyTableWithGeneratorIterator_Splits_Cancel
Date Mon, 01 Jun 2015 15:44:17 GMT
Josh Elser created ACCUMULO-3877:
------------------------------------

             Summary: TableOperationsIT failed in testCompactEmptyTableWithGeneratorIterator_Splits_Cancel
                 Key: ACCUMULO-3877
                 URL: https://issues.apache.org/jira/browse/ACCUMULO-3877
             Project: Accumulo
          Issue Type: Bug
          Components: test
    Affects Versions: 1.7.0
            Reporter: Josh Elser
            Assignee: Josh Elser
            Priority: Minor
             Fix For: 1.7.1, 1.8.0


{noformat}
Scan should be empty if compaction canceled. Actual is {a1 colF3:colQ3 [] 1432871214628 false=1,
c1 colF3:colQ3 [] 1432871214628 false=1}
{noformat}

It looks like the test is making an assumption that we'll see all of the entries this iterator
makes.

{code}

  static {
    SortedMap<Key,Value> t = new TreeMap<>();
    t.put(new Key(new Text("a1"), new Text("colF3"), new Text("colQ3"), System.currentTimeMillis()),
new Value("1".getBytes()));
    t.put(new Key(new Text("c1"), new Text("colF3"), new Text("colQ3"), System.currentTimeMillis()),
new Value("1".getBytes()));
    t.put(new Key(new Text("m1"), new Text("colF3"), new Text("colQ3"), System.currentTimeMillis()),
new Value("1".getBytes()));
    allEntriesToInject = Collections.unmodifiableSortedMap(t); // for safety
  }
{code}

However, because the split is added to the table prior to the compaction

{code}
splitset.add(new Text("f"));
{code}

It's possible that we see no entries, the "m1" key, the "a1" and "c1" keys, or all three keys.
The test doesn't consider the middle two cases, and expects all or none.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message