[
https://issues.apache.org/jira/browse/CASSANDRA-12183?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16818193#comment-16818193
]
Alex Deparvu commented on CASSANDRA-12183:
------------------------------------------
it seems the same generator code is used for the [original
uuid|https://github.com/apache/cassandra/blob/trunk/src/java/org/apache/cassandra/db/lifecycle/LogTransaction.java#L126],
so in theory there should not be a big issue if the existing id would be used
instead of generating a new one, but I don't really know the code so this might
be off.
I can volunteer a patch, but I can't do more than 'run all tests' to validate
this.
bq. I wonder if this is really the bug - should we be storing the UUID we print
out?
I agree it makes sense to use the same id on both places but I would rather see
this filed as an improvement, not a bug.
> compaction_history system table does not capture all historical compaction
> sessions
> -----------------------------------------------------------------------------------
>
> Key: CASSANDRA-12183
> URL: https://issues.apache.org/jira/browse/CASSANDRA-12183
> Project: Cassandra
> Issue Type: Bug
> Components: Legacy/Observability, Local/Compaction
> Reporter: Wei Deng
> Priority: Normal
>
> It appears that some compaction sessions are not recorded in
> system.compaction_history table after the compaction session successfully
> finishes.
> The following is an example (test by simply running +cassandra-stress write
> n=100000000+):
> {noformat}
> automaton@wdengdse50google-98425b985-3:~$ nodetool compactionstats
> pending tasks: 46
> id compaction type keyspace
> table completed total unit progress
> fa8a4f30-4884-11e6-b916-1dbd340a212f Compaction keyspace1
> standard1 4233184044 4774209875 bytes 88.67%
> 91e30d21-4887-11e6-b916-1dbd340a212f Compaction keyspace1
> standard1 836983029 889773060 bytes 94.07%
> Active compaction remaining time : 0h00m35s
> automaton@wdengdse50google-98425b985-3:~$ nodetool compactionstats
> pending tasks: 47
> id compaction type keyspace
> table completed total unit progress
> fa8a4f30-4884-11e6-b916-1dbd340a212f Compaction keyspace1
> standard1 4353251539 4774209875 bytes 91.18%
> 28359094-4888-11e6-b916-1dbd340a212f Compaction keyspace1
> standard1 49732274 4071652280 bytes 1.22%
> Active compaction remaining time : 0h04m24s
> {noformat}
> At this point you know the previous compaction session
> 91e30d21-4887-11e6-b916-1dbd340a212f finished and confirmation can be found
> from debug.log
> {noformat}
> automaton@wdengdse50google-98425b985-3:~$ grep
> 91e30d21-4887-11e6-b916-1dbd340a212f /var/log/cassandra/debug.log
> DEBUG [CompactionExecutor:4] 2016-07-12 23:22:58,674 CompactionTask.java:146
> - Compacting (91e30d21-4887-11e6-b916-1dbd340a212f)
> [/var/lib/cassandra/data/keyspace1/standard1-9c02e9c1487c11e6b9161dbd340a212f/mb-290-big-Data.db:level=0,
>
> /var/lib/cassandra/data/keyspace1/standard1-9c02e9c1487c11e6b9161dbd340a212f/mb-279-big-Data.db:level=0,
>
> /var/lib/cassandra/data/keyspace1/standard1-9c02e9c1487c11e6b9161dbd340a212f/mb-281-big-Data.db:level=0,
>
> /var/lib/cassandra/data/keyspace1/standard1-9c02e9c1487c11e6b9161dbd340a212f/mb-280-big-Data.db:level=0,
>
> /var/lib/cassandra/data/keyspace1/standard1-9c02e9c1487c11e6b9161dbd340a212f/mb-284-big-Data.db:level=0,
>
> /var/lib/cassandra/data/keyspace1/standard1-9c02e9c1487c11e6b9161dbd340a212f/mb-283-big-Data.db:level=0,
>
> /var/lib/cassandra/data/keyspace1/standard1-9c02e9c1487c11e6b9161dbd340a212f/mb-287-big-Data.db:level=0,
>
> /var/lib/cassandra/data/keyspace1/standard1-9c02e9c1487c11e6b9161dbd340a212f/mb-292-big-Data.db:level=0,
>
> /var/lib/cassandra/data/keyspace1/standard1-9c02e9c1487c11e6b9161dbd340a212f/mb-286-big-Data.db:level=0,
>
> /var/lib/cassandra/data/keyspace1/standard1-9c02e9c1487c11e6b9161dbd340a212f/mb-289-big-Data.db:level=0,
> ]
> DEBUG [CompactionExecutor:4] 2016-07-12 23:26:56,054 CompactionTask.java:217
> - Compacted (91e30d21-4887-11e6-b916-1dbd340a212f) 10 sstables to
> [/var/lib/cassandra/data/keyspace1/standard1-9c02e9c1487c11e6b9161dbd340a212f/mb-293-big,]
> to level=0. 889,773,060 bytes to 890,473,350 (~100% of original) in
> 237,365ms = 3.577703MB/s. 0 total partitions merged to 3,871,921. Partition
> merge counts were {1:3871921, }
> {noformat}
> However, if you query system.compaction_history table or run "nodetool
> compactionhistory | grep 91e30d21-4887-11e6-b916-1dbd340a212f" you will get
> nothing:
> {noformat}
> automaton@wdengdse50google-98425b985-3:~$ cqlsh -u cassandra
> Password:
> Connected to dse50 at 127.0.0.1:9042.
> [cqlsh 5.0.1 | Cassandra 3.0.7.1158 | DSE 5.0.0 | CQL spec 3.4.0 | Native
> protocol v4]
> Use HELP for help.
> cassandra@cqlsh> select * from system.compaction_history where
> id=91e30d21-4887-11e6-b916-1dbd340a212f;
> id | bytes_in | bytes_out | columnfamily_name | compacted_at | keyspace_name
> | rows_merged
> ----+----------+-----------+-------------------+--------------+---------------+-------------
> (0 rows)
> automaton@wdengdse50google-98425b985-3:~$ nodetool flush system
> automaton@wdengdse50google-98425b985-3:~$ nodetool compactionhistory | grep
> 91e30d21-4887-11e6-b916-1dbd340a212f
> {noformat}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
