QianBase 运维实用命令

1. 统计所有启动ESP的MXOSRVR进程及ESP个数

sqps | grep esp| awk '{print $7}' | sort | uniq -c | sort -nk1

示例输出：

      4 $Z000C000EQQ
     10 $Z0007005FFT
     10 $Z000P00035N
     10 $Z000P0020DJ
     12 $Z000P002LWR
     12 $Z000P0046JZ
     16 $Z000800487E
     20 $Z0000004J6N
     20 $Z0000005VTS
     20 $Z0001000QCJ
     20 $Z0001000SMD
     20 $Z0001001NRV
     20 $Z0001002DJW
     20 $Z0001004M98
     20 $Z0001004QB8
     20 $Z0001005LR1
     20 $Z0001006YLX
     20 $Z00020030GX
     20 $Z0002005HJH
     20 $Z0003000DMH
     20 $Z0003000TNS

2. 统计当前有多少个MXOSRVR启动ESP进程

sqps | grep esp| awk '{print $7}' | sort | uniq | wc -l

3. 统计当前有多少个ESP进程

sqps | grep esp| awk '{print $6}' | sort | uniq | wc -l

4. 通过REST API获取连接信息

curl  -X GET -H "Accept: application/json" https://xx.xx.xx.xx:4201/v1/servers/dcs/connections -k

5. 查看哪些ESP所在的MXOSRVR处于连接状态

curl  -X GET -H "Accept: application/json" https://xx.xx.xx.xx:4201/v1/servers/dcs/connections -k > connections.log
sed 's/,{/,\n{/g' connections.log > connections1.log
sed -n 's/.*"STATE":"\(.*\)","NID".*"PROCESS_NAME":"\(.*\)","IP_ADDRESS".*/\1 \2/p' connections1.log | grep CONNECTED | awk '{print $2}' | sort | uniq > connected_mxosrvr.log
sqps | grep esp| awk '{print $7}' | sort | uniq > hasesps_mxosrvr.log
sort hasesps_mxosrvr.log connected_mxosrvr.log | uniq -d > hasesps_connected_mxosrvr.log

6. 查看各个节点RegionServer的线程数

edb_pdsh -a "ps -ef | grep 'RegionServer start' | grep -v 'grep' | awk -F ' ' '{print \$2}' | xargs top -b -n 1 -Hp | grep 'Threads'"

7. 指定连接到固定的MXOSRVR

方法一：
export _JAVA_OPTIONS="-Dt4jdbc.specifiedServer=qianbase1.esgyn.cn:23402" 

方法二：
trafci.sh -h qianbase1:23400/:specifiedServer=qianbase1.esgyn.cn:23402 -u trafodion -p traf123

8. 查看数据库启动时间

grep monitor monitor.map.nodename| grep BEGIN

9. 收集数据库所有日志及HBase日志

cds
./getlogs.sh --hbaselogs

10. 使用HBase ExportSnapshot接口备份恢复表

//备份
snapshot 'TRAF_RSRVD_5:TRAFODION._DTM_.TRAFODION_BINLOG', '36_BINLOG_20220707'
hbase org.apache.hadoop.hbase.snapshot.ExportSnapshot -snapshot 36_BINLOG_20220707 -copy-from hdfs://10.15.40.39:8020/hbase -copy-to hdfs://xx.xx.xx.xx:8020/hbase -mappers 32 -bandwidth 16384
//恢复
disable 'TRAF_RSRVD_5:TRAFODION._DTM_.TRAFODION_BINLOG'
drop 'TRAF_RSRVD_5:TRAFODION._DTM_.TRAFODION_BINLOG'
restore_snapshot '36_BINLOG_20220707'

11. 查看表的创建修改时间

SELECT  converttimestamp(create_time)+interval '8' HOUR AS create_time , 
        converttimestamp(redef_time)+interval '8' HOUR AS redef_time
FROM "_MD_".objects AS a WHERE a.object_name=<object_name>;

12. 修改表的memstore flush size大小

alter table <tablename>  alter hbase_options (MEMSTORE_FLUSH_SIZE='268435456');

13. 查看数据库中发生8102的表及rowkey信息

edb_pdsh -a "cd /var/log/hbase; grep -e 'checkAndPut EXIT - returns' -e 'checkAndPutRegionTx EXIT - returns' *" > /tmp/all.log
sed -n 's/.*in region \(.*TRAFODION.AFA.*\),.*,.*,skey.*/\1/p' /tmp/all.log | sort | uniq

14. 清除mxosrvr内的querycache及NATable cache

//清除QueryCache
cqd query_cache ‘0’;

//清除NATable Cache
方法1：cqd metadata_cache_size ‘0’;
方法2：delete all from table(natablecache('remove'));

15. 导出QueryCache

generate user querycache export '/opt/trafodion/cachefile.sql';

16. 统计所有ESP的启动时间及持续运行时间

edb_pdsh -a "ps -ef | grep tdm_arkesp | grep -v grep | awk '{print \$2}' | xargs ps -o lstart,etime -p"