生产实践--gbase读写有kerberos认证的hdfs

202 203配置读hdfs
show variables like '%keytab%';

按要求配置GC&GN节点--星环TDH提供
/etc/hosts配置所有节点

/etc/krb5.conf配置所有节点包括gc&gn，放到etc目录下，属主为root，权限644；--星环提供
 

3、hdfs新建用户gbase，授权，给过来的是gbase.keytab，改成gbase.kt；--星环创建用户并赋权
gbase.kt放到gc&gn的conf下面，属主是gbase；
GC目录为$GCLUSTER_BASE/config，GN目录为$GBASE_BASE/config；

show variables like 'gbase_hdfs%';
gccli在线设置，gn也会生效；
set global gbase_hdfs_auth_mode=Kerberos;
set global gbase_hdfs_namenodes='p-tdhhh5-mg2:50070,p-tdhhh9-mg3:50070';
set global gbase_hdfs_principal='gbase@TDH';
(gbase用户密码abc@123)

在线设置参数 要退出当前实例 重新进才能看到

torc表测试：
建表：string改成varchar(200)  
PARTITIONED BY改成PARTITION BY
BY RANGE (partition_date) ：()里面只能有字段，不能有字段注释
表的COMMENT不能放在最后

前5个字段过滤掉，最后面加一个 autofill

use vc_ggjc.readhdfs;
load data infile 'hdfs://p-tdhhh5-mg2:8020/inceptor1/user/hive/warehouse/ads_rcdb.db/ads-rcdb/s_agr_loan_inr_acct_info/p202310/base_101382618/*' into table s_agr_loan_inr_acct_info data_format orc;
--报错

可以*/* ，精确到文件
load data infile 'hdfs://p-tdhhh5-mg2:8020/inceptor1/user/hive/warehouse/ads_rcdb.db/ads-rcdb/s_agr_loan_inr_acct_info/p202310/*/*' into table s_agr_loan_inr_acct_info_tmp20240108 
fields terminated by '|' table_fields 't1 filler,t2 filler,t3 filler,t4 filler,t5 filler,data_src,data_dt,dubil_no,contr_no,cust_no,prdt_no,src_prdt_no,src_prdt_name,subject_code,curncy_cd,
dubil_amt,dubil_blnc,contr_amt,contr_blnc,crdt_limit_amt,crdt_limit_used,crdt_limit_rest,putout_dt,matrty_dt,loan_term_d,loan_term_m,loan_term_rest,base_rate,exct_rate,overd_rate,
loan_acct_no,repay_acct_no,loan_purpose,loan_form_cd,guarn_way_cd,direct_indust_cd,loan_status_cd,five_class_cd,payment_mode_cd,repay_mode_cd,int_calcu_mode_cd,sttl_int_cycle_cd,adjust_mode_cd,
flag_agricul,flag_low_risk,is_govrmt_fin_pltfm,flag_settle,settle_dt,flag_extend,extend_times,flag_charge_off,charge_off_dt,charge_off_amt,charge_off_intrst,extract_ready_val,flag_overd,overd_amt,
overd_dt,overd_prncpl_amt,overd_prncpl_dt,overd_intrst_amt,overd_intrst_dt,overd_times_sum,overd_times_max,int_debt_blnc,ext_debt_blnc,normal_bal,overdue_bal,sluggish_bal,doubtful_bal,
prior_repay_dt,prior_repay_amt,prior_actual_amt,prior_repay_state_cd,next_repay_dt,next_repay_amt,fina_orgn_no,oper_orgn_no,oper_user_no,crt_no,manage_org_no,src_subject_code,maturity,prdt_one_no,
prdt_one_name,relative_agreement,refinanc_time,fund_use_loc_cd,acct_typ,agriculture_typ,shld_pay_prncpl,shld_pay_intrst,actl_pay_prncpl,actl_pay_intrst,loan_type_cd,strong_guarantee_typ,
partition_date,frst_rate,is_online,propertyloan_typ,cust_name,no_rep_prin_typ,online_typ'   data_format orc;

加载text表
use vc_ggjc.readhdfs;
load data infile 'hdfs://p-tdhhh5-mg2:8020/inceptor1/user/hive/warehouse/ads_rcdb.db/ads-rcdb/s_prt_indv_acct_blnc_chg/base_100493312/*' into table s_prt_indv_acct_blnc_chg1025 data_format 3;

加载orc表
use vc_ggjc.readhdfs;
load data infile 'hdfs://p-tdhhh5-mg2:8020/inceptor1/user/hive/warehouse/ads_rcdb.db/ads-rcdb/s_prt_indv_acct_blnc_chg/base_100493312/*' into table s_prt_indv_acct_blnc_chg1025 data_format orc;