You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@carbondata.apache.org by "Ajeet Rai (JIRA)" <ji...@apache.org> on 2018/02/06 13:08:00 UTC

[jira] [Closed] (CARBONDATA-1761) (Carbon1.3.0 - DELETE SEGMENT BY ID) In Progress Segment is marked for delete if respective id is given in delete segment by id query

     [ https://issues.apache.org/jira/browse/CARBONDATA-1761?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Ajeet Rai closed CARBONDATA-1761.
---------------------------------

Issue is fixed. A segment is not deleted if load is still in progress for that segment id.

> (Carbon1.3.0 - DELETE SEGMENT BY ID) In Progress Segment is marked for delete if respective id is given in delete segment by id query
> -------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: CARBONDATA-1761
>                 URL: https://issues.apache.org/jira/browse/CARBONDATA-1761
>             Project: CarbonData
>          Issue Type: Bug
>          Components: data-load
>    Affects Versions: 1.3.0
>         Environment: 3 Node ant cluster
> Description
>            Reporter: Ajeet Rai
>            Assignee: Akash R Nilugal
>            Priority: Major
>              Labels: dfx
>             Fix For: 1.3.0
>
>          Time Spent: 4h 20m
>  Remaining Estimate: 0h
>
> (Carbon1.3.0 - DELETE SEGMENT BY ID) In Progress Segment is marked for delete if respective id is given in delete segment by id query.
> 1: Create a table
> CREATE TABLE IF NOT EXISTS flow_carbon_new999(txn_dte String,dt String,txn_bk String,txn_br String,own_bk String,own_br String,opp_bk String,bus_opr_cde String,opt_prd_cde String,cus_no String,cus_ac String,opp_ac_nme String,opp_ac String,bv_no String,aco_ac String,ac_dte String,txn_cnt int,jrn_par int,mfm_jrn_no String,cbn_jrn_no String,ibs_jrn_no String,vch_no String,vch_seq String,srv_cde String,bus_cd_no String,id_flg String,bv_cde String,txn_time String,txn_tlr String,ety_tlr String,ety_bk String,ety_br String,bus_pss_no String,chk_flg String,chk_tlr String,chk_jrn_no String, bus_sys_no String,txn_sub_cde String,fin_bus_cde String,fin_bus_sub_cde String,chl String,tml_id String,sus_no String,sus_seq String, cho_seq String, itm_itm String,itm_sub String,itm_sss String,dc_flg String,amt decimal(15,2),bal decimal(15,2),ccy String,spv_flg String,vch_vld_dte String,pst_bk String,pst_br String,ec_flg String,aco_tlr String,gen_flg String,his_rec_sum_flg String,his_flg String,vch_typ String,val_dte String,opp_ac_flg String,cmb_flg String,ass_vch_flg String,cus_pps_flg String,bus_rmk_cde String,vch_bus_rmk String,tec_rmk_cde String,vch_tec_rmk String,gems_last_upd_d String,maps_date String,maps_job String)STORED BY 'org.apache.carbondata.format' TBLPROPERTIES('DICTIONARY_INCLUDE'='txn_cnt,jrn_par,amt,bal','No_Inverted_Index'= 'txn_dte,dt,txn_bk,txn_br,own_bk ,own_br ,opp_bk ,bus_opr_cde ,opt_prd_cde ,cus_no ,cus_ac ,opp_ac_nme ,opp_ac ,bv_no ,aco_ac ,ac_dte ,txn_cnt ,jrn_par ,mfm_jrn_no ,cbn_jrn_no ,ibs_jrn_no ,vch_no ,vch_seq ,srv_cde ,bus_cd_no ,id_flg ,bv_cde ,txn_time ,txn_tlr ,ety_tlr ,ety_bk ,ety_br ,bus_pss_no ,chk_flg ,chk_tlr ,chk_jrn_no , bus_sys_no ,txn_sub_cde ,fin_bus_cde ,fin_bus_sub_cde ,chl ,tml_id ,sus_no ,sus_seq , cho_seq , itm_itm ,itm_sub ,itm_sss ,dc_flg ,amt,bal,ccy ,spv_flg ,vch_vld_dte ,pst_bk ,pst_br ,ec_flg ,aco_tlr ,gen_flg ,his_rec_sum_flg ,his_flg ,vch_typ ,val_dte ,opp_ac_flg ,cmb_flg ,ass_vch_flg ,cus_pps_flg ,bus_rmk_cde ,vch_bus_rmk ,tec_rmk_cde ,vch_tec_rmk ,gems_last_upd_d ,maps_date ,maps_job' );
> 2: start a data load.
> LOAD DATA inpath 'hdfs://hacluster/user/test/20140101_1_144444444.csv' into table flow_carbon_new999 options('DELIMITER'=',', 'QUOTECHAR'='"','header'='false');
> 3: run a insert into/overwrite job
> insert into table flow_carbon_new999 select * from flow_carbon_new666;
> 4: show segments for table flow_carbon_new999;
> 5: Observe that load/insert/overwrite job is started with new segment id
> 6: now run a delete segment by id query with this id.
> DELETE FROM TABLE ajeet.flow_carbon_new999 WHERE SEGMENT.ID IN (34)
> 7: again run show segment and see this segment which is still in progress is marked for delete.
> 8: Observe that insert/load job is still running and after some time(in next job of load/insert/overwrite), this job fails with below error:
> Error: java.lang.RuntimeException: It seems insert overwrite has been issued during load (state=,code=0)
> This is not correct behaviour and it should be handled.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)