[
https://issues.apache.org/jira/browse/CASSANDRA-7019?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15154056#comment-15154056
]
Branimir Lambov commented on CASSANDRA-7019:
--------------------------------------------
Results of the longer run last evening:
{code}
Hashes: [495354, 748895267, -2068352751, 1145481084, -1976889921, 9, 2539,
-34], retrieved in 17.772s
{"provide_overlapping_tombstones":"CELL","class":"org.apache.cassandra.db.compaction.LeveledCompactionStrategy","sstable_size_in_mb":"16"}
CELL compactions completed in 13.662s
Operations completed in 3147.977s, out of which 88.668 for ongoing CELL
background compactions
At start: 53 tables 1015895699 bytes 985074 rows 450179
deleted rows 49132 tombstone markers
At end: 53 tables 522294953 bytes 526978 rows 274600
deleted rows 35211 tombstone markers
Hashes: [495354, 748895267, -2068352751, 1145481084, -1976889921, 9, 2539,
-34], retrieved in 16.924s
Hashes: [495354, 748895267, -2068352751, 1145481084, -1976889921, 9, 2539,
-34], retrieved in 18.580s
{"provide_overlapping_tombstones":"ROW","class":"org.apache.cassandra.db.compaction.LeveledCompactionStrategy","sstable_size_in_mb":"16"}
ROW compactions completed in 11.817s
Operations completed in 3370.982s, out of which 87.959 for ongoing ROW
background compactions
At start: 52 tables 1034287909 bytes 983095 rows 444773
deleted rows 48188 tombstone markers
At end: 52 tables 555226285 bytes 527258 rows 274600
deleted rows 35055 tombstone markers
Hashes: [495354, 748895267, -2068352751, 1145481084, -1976889921, 9, 2539,
-34], retrieved in 15.539s
Hashes: [495354, 748895267, -2068352751, 1145481084, -1976889921, 9, 2539,
-34], retrieved in 17.840s
{"provide_overlapping_tombstones":"NONE","class":"org.apache.cassandra.db.compaction.LeveledCompactionStrategy","sstable_size_in_mb":"16"}
NONE compactions completed in 17.979s
Operations completed in 4319.732s, out of which 83.286 for ongoing NONE
background compactions
At start: 57 tables 1333377280 bytes 1267545 rows 556267
deleted rows 49755 tombstone markers
At end: 57 tables 1333372458 bytes 1267545 rows 556267
deleted rows 49755 tombstone markers
Hashes: [495354, 748895267, -2068352751, 1145481084, -1976889921, 9, 2539,
-34], retrieved in 17.709s
Hashes: [495354, 748895267, -2068352751, 1145481084, -1976889921, 9, 2539,
-34], retrieved in 19.713s
{"max_threshold":"32","min_threshold":"4","provide_overlapping_tombstones":"CELL","min_sstable_size":"0","class":"org.apache.cassandra.db.compaction.SizeTieredCompactionStrategy"}
CELL compactions completed in 24.677s
Operations completed in 4421.714s, out of which 91.985 for ongoing CELL
background compactions
At start: 23 tables 1842748717 bytes 1780330 rows 465835
deleted rows 52040 tombstone markers
At end: 23 tables 522233965 bytes 542078 rows 274600
deleted rows 34771 tombstone markers
Hashes: [495354, 748895267, -2068352751, 1145481084, -1976889921, 9, 2539,
-34], retrieved in 17.628s
Hashes: [495354, 748895267, -2068352751, 1145481084, -1976889921, 9, 2539,
-34], retrieved in 19.095s
{"max_threshold":"32","min_threshold":"4","provide_overlapping_tombstones":"ROW","min_sstable_size":"0","class":"org.apache.cassandra.db.compaction.SizeTieredCompactionStrategy"}
ROW compactions completed in 12.763s
Operations completed in 4223.531s, out of which 59.436 for ongoing ROW
background compactions
At start: 25 tables 1744662122 bytes 1662199 rows 526525
deleted rows 63241 tombstone markers
At end: 25 tables 566431602 bytes 538314 rows 274600
deleted rows 37604 tombstone markers
Hashes: [495354, 748895267, -2068352751, 1145481084, -1976889921, 9, 2539,
-34], retrieved in 15.674s
Hashes: [495354, 748895267, -2068352751, 1145481084, -1976889921, 9, 2539,
-34], retrieved in 20.745s
{"max_threshold":"32","min_threshold":"4","provide_overlapping_tombstones":"NONE","min_sstable_size":"0","class":"org.apache.cassandra.db.compaction.SizeTieredCompactionStrategy"}
NONE compactions completed in 18.571s
Operations completed in 8617.601s, out of which 79.785 for ongoing NONE
background compactions
At start: 20 tables 2939870504 bytes 2800985 rows 756021
deleted rows 69308 tombstone markers
At end: 20 tables 2939870070 bytes 2800985 rows 756021
deleted rows 69308 tombstone markers
Hashes: [495354, 748895267, -2068352751, 1145481084, -1976889921, 9, 2539,
-34], retrieved in 21.108s
{code}
There are two differences here:
- The test now runs hash UDF/UDAs before and after the GC compaction at the end
and compares the results to verify no data has changed.
- The compactions applied during the construction of the test data ("background
compactions" above) are of the tested type.
The branch is updated with the new version of the test.
> Improve tombstone compactions
> -----------------------------
>
> Key: CASSANDRA-7019
> URL: https://issues.apache.org/jira/browse/CASSANDRA-7019
> Project: Cassandra
> Issue Type: Improvement
> Components: Compaction
> Reporter: Marcus Eriksson
> Assignee: Branimir Lambov
> Labels: compaction
> Fix For: 3.x
>
>
> When there are no other compactions to do, we trigger a single-sstable
> compaction if there is more than X% droppable tombstones in the sstable.
> In this ticket we should try to include overlapping sstables in those
> compactions to be able to actually drop the tombstones. Might only be doable
> with LCS (with STCS we would probably end up including all sstables)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
