[
https://issues.apache.org/jira/browse/GEODE-8958?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17290527#comment-17290527
]
ASF GitHub Bot commented on GEODE-8958:
---------------------------------------
jchen21 commented on a change in pull request #6042:
URL: https://github.com/apache/geode/pull/6042#discussion_r582366221
##########
File path:
geode-core/src/distributedTest/java/org/apache/geode/internal/cache/versions/TombstoneDUnitTest.java
##########
@@ -117,35 +118,74 @@ public void
testTombstoneGcMessagesBetweenPersistentAndNonPersistentRegion() {
});
}
+
@Test
- public void testGetOldestTombstoneTimeReplicate() {
+ public void
testWhenAnOutOfRangeTimeStampIsSeenWeExpireItInReplicateTombstoneSweeper() {
VM server1 = VM.getVM(0);
VM server2 = VM.getVM(1);
+ final int FAR_INTO_THE_FUTURE = 1000000; // 1 million millis into the
future
+ final int count = 10;
server1.invoke(() -> {
- createCacheAndRegion(REPLICATE_PERSISTENT);
- region.put("K1", "V1");
- region.put("K2", "V2");
+ createCacheAndRegion(RegionShortcut.REPLICATE_PERSISTENT);
+ for (int i = 0; i < count; i++) {
+ region.put("K" + i, "V" + i);
+ }
});
- server2.invoke(() -> createCacheAndRegion(REPLICATE));
+ server2.invoke(() -> createCacheAndRegion(RegionShortcut.REPLICATE));
server1.invoke(() -> {
+ TombstoneService.TombstoneSweeper tombstoneSweeper =
+ ((InternalCache)
cache).getTombstoneService().getSweeper((LocalRegion) region);
+
+ RegionEntry regionEntry = ((LocalRegion) region).getRegionEntry("K0");
+ VersionTag<?> versionTag = regionEntry.getVersionStamp().asVersionTag();
+ versionTag.setVersionTimeStamp(System.currentTimeMillis() +
FAR_INTO_THE_FUTURE);
+
+ TombstoneService.Tombstone modifiedTombstone =
+ new TombstoneService.Tombstone(regionEntry, (LocalRegion) region,
+ versionTag);
+ tombstoneSweeper.tombstones.add(modifiedTombstone);
+ tombstoneSweeper.checkOldestUnexpired(System.currentTimeMillis());
// Send tombstone gc message to vm1.
- region.destroy("K1");
+ assertThat(tombstoneSweeper.getOldestTombstoneTime()).isEqualTo(0);
+ });
+ }
+ @Test
+ public void testGetOldestTombstoneTimeForReplicateTombstoneSweeper() {
+ VM server1 = VM.getVM(0);
+ VM server2 = VM.getVM(1);
+ final int count = 10;
+ server1.invoke(() -> {
+ createCacheAndRegion(RegionShortcut.REPLICATE_PERSISTENT);
+ for (int i = 0; i < count; i++) {
+ region.put("K" + i, "V" + i);
+ }
+ });
+
+ server2.invoke(() -> createCacheAndRegion(RegionShortcut.REPLICATE));
+
+ server1.invoke(() -> {
TombstoneService.TombstoneSweeper tombstoneSweeper =
((InternalCache)
cache).getTombstoneService().getSweeper((LocalRegion) region);
+ // Send tombstone gc message to vm1.
+ for (int i = 0; i < count; i++) {
+ region.destroy("K" + i);
+ assertThat(
Review comment:
First, usually immediately after `Region.destroy()`, the difference
between `tombstoneSweeper.getOldestTombstoneTime()` and
`System.currentTimeMillis()` is only a few milliseconds. So asserting that
```
tombstoneSweeper.getOldestTombstoneTime()
+ TombstoneService.REPLICATE_TOMBSTONE_TIMEOUT_DEFAULT -
System.currentTimeMillis())
.isGreaterThan(0)
```
is like asserting `REPLICATE_TOMBSTONE_TIMEOUT_DEFAULT` is greater than 0,
i.e. asserting 600000 > 0. There is nothing wrong with that. However, a better
assertion would be the one you deleted:
```
assertThat(tombstoneSweeper.getOldestTombstoneTime()).isGreaterThan(0)
.isLessThan(((InternalCache) cache).cacheTimeMillis());
```
Second, if we expect that GC clears out tombstones, I would recommend moving
the following assertion
```
assertThat(tombstoneSweeper.getOldestTombstoneTime()).isEqualTo(0)
```
into the `for` loop, instead of after the `for` loop.
Please correct me if I am wrong.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
> Tombstone expiration: in the event that a version timestamp is too far in the
> future, expire the tombstone
> ----------------------------------------------------------------------------------------------------------
>
> Key: GEODE-8958
> URL: https://issues.apache.org/jira/browse/GEODE-8958
> Project: Geode
> Issue Type: Bug
> Components: core
> Reporter: Mark Hanson
> Assignee: Mark Hanson
> Priority: Major
> Labels: pull-request-available
>
> We are seeing a bug where for some reason, the version timestamp on the
> tombstone is too far into the future to be realistic.
>
> In such a case, we are going to expire the tombstone as soon as we see it.
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
