hbase 删除数据 释放空间
2024-06-28 16:00:58
供稿：网友
# hbasecheck hbase sizehdfs dfs -du -h /hbase/data/default/[root@ZHJT machtalk]# hdfs dfs -du -h /hbase/data/default197.4 M 197.4 M /hbase/data/default/API_access_RECORD1.0 M 1.0 M /hbase/data/default/DEV_OPT38.6 M 38.6 M /hbase/data/default/LOG_RECORD333 333 /hbase/data/default/dev_fault13.4 M 13.4 M /hbase/data/default/value_datadescribe 'API_ACCESS_RECORD'alter 'tableName′, NAME => 'API_ACCESS_RECORD', TTL => 2592000alter 'API_ACCESS_RECORD', NAME => 'values', TTL => 259200[machtalk@ZHJT ~]$ df Filesystem 1K-blocks Used Available Use% Mounted on/dev/mapper/vg_test-lv_root 51606140 24709820 24274880 51% /tmpfs 30692756 8 30692748 1% /dev/shm/dev/sda1 495844 32422 437822 7% /boot/dev/mapper/vg_test-lv_home 461914096 33595976 404854184 8% /datacm_PRocesses 30692756 10616 30682140 1% /opt/cm-5.5.0/run/cloudera-scm-agent/processdisable 'table' enable 'test_ttl' get 'API_ACCESS_RECORD','values'# 修改alter 'API_ACCESS_RECORD', NAME => 'values', TTL => 2592000# 修改过程hbase(main):005:0> alter 'API_ACCESS_RECORD', NAME => 'values', TTL => 259200Updating all regions with the new schema...0/1 regions updated.1/1 regions updated.Done.0 row(s) in 2.5270 seconds# 再次查看 describe 'API_ACCESS_RECORD'describe 'API_ACCESS_RECORD'[machtalk@ZHJT ~]$ df Filesystem 1K-blocks Used Available Use% Mounted on/dev/mapper/vg_test-lv_root 51606140 24713260 24271440 51% /tmpfs 30692756 8 30692748 1% /dev/shm/dev/sda1 495844 32422 437822 7% /boot/dev/mapper/vg_test-lv_home 461914096 33581168 404868992 8% /datacm_processes 30692756 10616 30682140 1% /opt/cm-5.5.0/run/cloudera-scm-agent/process[machtalk@ZHJT ~]$ hdfs dfs -du -h /hbase/data/default/197.4 M 197.4 M /hbase/data/default/API_ACCESS_RECORD1.0 M 1.0 M /hbase/data/default/DEV_OPT38.6 M 38.6 M /hbase/data/default/LOG_RECORD333 333 /hbase/data/default/dev_fault13.4 M 13.4 M /hbase/data/default/value_datat1 = get_table 'API_ACCESS_RECORD'hbase(main):017:0> t1._count_internal=> 86556# 修改表的默认大小alter 't1', METHOD => 'table_att', *MAX_FILESIZE* => '134217728' Maximum HStoreFile size. If any one of a column families' HStoreFiles has grown to exceed this value, the hosting HRegion is split in two.# 参考http://blog.csdn.net/mrtitan/article/details/82920411480521600000 1483200000000# 1月3日之前都没有数据scan 'API_ACCESS_RECORD', {TIMERANGE =>[1480521600000,1483372800000]}# 保存十天试试alter 'API_ACCESS_RECORD', NAME => 'values', TTL => 864000# 删除完成后再看hbase(main):006:0> t1._count_internal=> 4148[machtalk@ZHJT ~]$ df Filesystem 1K-blocks Used Available Use% Mounted on/dev/mapper/vg_test-lv_root 51606140 24721924 24262776 51% /tmpfs 30692756 8 30692748 1% /dev/shm/dev/sda1 495844 32422 437822 7% /boot/dev/mapper/vg_test-lv_home 461914096 33614808 404835352 8% /datacm_processes 30692756 10616 30682140 1% /opt/cm-5.5.0/run/cloudera-scm-agent/process24721924 - 24709820 = 12104 byte# 只有大合并之后，空间才会释放https://dxer.github.io/2016/03/18/hbase/http://book.51cto.com/art/201312/420269.htm# 大合并，释放空间hbase shellmajor_compact[machtalk@ZHJT ~]$ df Filesystem 1K-blocks Used Available Use% Mounted on/dev/mapper/vg_test-lv_root 51606140 24730240 24254460 51% /tmpfs 30692756 8 30692748 1% /dev/shm/dev/sda1 495844 32422 437822 7% /boot/dev/mapper/vg_test-lv_home 461914096 33607744 404842416 8% /datacm_processes 30692756 10616 30682140 1% /opt/cm-5.5.0/run/cloudera-scm-agent/processhttp://www.CUOXin.com/nexiyi/p/hbase_shell.html# 应该hadoop这块没有释放 hdfshdfs dfs -du -h /hbase/.Trash
