[ https://issues.apache.org/jira/browse/CARBONDATA-4034?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Jiayu Shen updated CARBONDATA-4034: ----------------------------------- Description: In the update flow, horizontal compaction will be significantly slower when updating with a lot of segments(or a lot of blocks). There is a case whose costing is as shown in the log. 2020-10-10 09:38:10,466 | INFO | [OperationManager-Background-Pool-28] | Horizontal Update Compaction operation started for [ods_oms.oms_wh_outbound_order] 2020-10-10 09:50:25,718 | INFO | [OperationManager-Background-Pool-28] | Horizontal Update Compaction operation completed for [ods_oms.oms_wh_outbound_order]. 2020-10-10 10:15:44,302 | INFO | [OperationManager-Background-Pool-28] | Horizontal Delete Compaction operation started for [ods_oms.oms_wh_outbound_order] 2020-10-10 10:15:54,874 | INFO | [OperationManager-Background-Pool-28] | Horizontal Delete Compaction operation completed for [ods_oms.oms_wh_outbound_order]. In this PR, we optimize the process between second and third row of the log, by optimizing the method _performDeleteDeltaCompaction_ in horizontal compaction flow. was: In the update flow, horizontal compaction will be significantly slower when updating with a lot of segments(or a lot of blocks). There is a case whose costing is as shown in the log. 2020-10-10 09:38:10,466 | INFO | [OperationManager-Background-Pool-28] | Horizontal Update Compaction operation started for [ods_oms.oms_wh_outbound_order] 2020-10-10 09:50:25,718 | INFO | [OperationManager-Background-Pool-28] | Horizontal Update Compaction operation completed for [ods_oms.oms_wh_outbound_order]. 2020-10-10 10:15:44,302 | INFO | [OperationManager-Background-Pool-28] | Horizontal Delete Compaction operation started for [ods_oms.oms_wh_outbound_order] 2020-10-10 10:15:54,874 | INFO | [OperationManager-Background-Pool-28] | Horizontal Delete Compaction operation completed for [ods_oms.oms_wh_outbound_order]. In this PR, we optimize the process between second and third row of the log, by optimizing the method _performDeleteDeltaCompaction_ in horizontal compaction flow. > Improve the time-comsuming of Horizontal Compaction for update > -------------------------------------------------------------- > > Key: CARBONDATA-4034 > URL: https://issues.apache.org/jira/browse/CARBONDATA-4034 > Project: CarbonData > Issue Type: Bug > Reporter: Jiayu Shen > Priority: Minor > > In the update flow, horizontal compaction will be significantly slower when updating with a lot of segments(or a lot of blocks). There is a case whose costing is as shown in the log. > 2020-10-10 09:38:10,466 | INFO | [OperationManager-Background-Pool-28] | Horizontal Update Compaction operation started for [ods_oms.oms_wh_outbound_order] > 2020-10-10 09:50:25,718 | INFO | [OperationManager-Background-Pool-28] | Horizontal Update Compaction operation completed for [ods_oms.oms_wh_outbound_order]. > 2020-10-10 10:15:44,302 | INFO | [OperationManager-Background-Pool-28] | Horizontal Delete Compaction operation started for [ods_oms.oms_wh_outbound_order] > 2020-10-10 10:15:54,874 | INFO | [OperationManager-Background-Pool-28] | Horizontal Delete Compaction operation completed for [ods_oms.oms_wh_outbound_order]. > In this PR, we optimize the process between second and third row of the log, by optimizing the method _performDeleteDeltaCompaction_ in horizontal compaction flow. > -- This message was sent by Atlassian Jira (v8.3.4#803005) |
Free forum by Nabble | Edit this page |