You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@iotdb.apache.org by "ZhangZhengming (Jira)" <ji...@apache.org> on 2022/11/09 10:03:00 UTC
[jira] [Assigned] (IOTDB-4895) [0.14][cluster][compaction] unsequence compaction lose 40% of data
[ https://issues.apache.org/jira/browse/IOTDB-4895?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
ZhangZhengming reassigned IOTDB-4895:
-------------------------------------
Assignee: (was: Liuxuxin)
> [0.14][cluster][compaction] unsequence compaction lose 40% of data
> ------------------------------------------------------------------
>
> Key: IOTDB-4895
> URL: https://issues.apache.org/jira/browse/IOTDB-4895
> Project: Apache IoTDB
> Issue Type: Bug
> Reporter: ZhangZhengming
> Priority: Major
> Attachments: image-2022-11-09-17-53-33-923.png, image-2022-11-09-17-55-58-042.png, image-2022-11-09-17-59-25-165.png
>
>
> I want to achieve quick merge by modifying configuration parameters.
> node: 1C3D, factor=1, multileader
> compaction para:
> {code:java}
> # compaction
> enable_seq_space_compaction=true
> enable_unseq_space_compaction=true
> enable_cross_space_compaction=true
> max_inner_compaction_candidate_file_num=2
> compaction_schedule_interval_in_ms=1000
> compaction_submission_interval_in_ms=1000
> compaction_write_throughput_mb_per_sec=0{code}
>
> benchmark para:
> {code:java}
> GROUP_NUMBER=30
> DEVICE_NUMBER=300
> SENSOR_NUMBER=1000
> IS_OUT_OF_ORDER=true
> OUT_OF_ORDER_RATIO=0.5
> CLIENT_NUMBER=30
> LOOP=1000
> BATCH_SIZE_PER_WRITE=100 {code}
>
> Here are the results of benchmark:
> !image-2022-11-09-17-53-33-923.png|width=830,height=245!
>
> In fact.
> !image-2022-11-09-17-55-58-042.png|width=758,height=266!
>
> Use third-party tools, the amount of data lost is different, but approximate
> !image-2022-11-09-17-59-25-165.png|width=391,height=379!
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)