说明
starrocks 集群的 be 节点的多个存储目录，有一个磁盘空间过小，搭建时误配置上，如下所示：
$ df -lh
Filesystem Size Used Avail Use% Mounted on
devtmpfs 94G 0 94G 0% /dev
tmpfs 94G 0 94G 0% /dev/shm
tmpfs 94G 43M 94G 1% /run
tmpfs 94G 0 94G 0% /sys/fs/cgroup
/dev/mapper/rootvg-lv_root 127G 1.7G 126G 2% /
/dev/mapper/rootvg-lv_usr 97G 1.9G 96G 2% /usr
/dev/sda2 506M 226M 281M 45% /boot
/dev/sda1 200M 12M 189M 6% /boot/efi
/dev/sdc 1.8T 61G 1.7T 4% /data19
/dev/sdo 7.3T 59G 7.3T 1% /data12
/dev/sdn 7.3T 62G 7.3T 1% /data11
/dev/sdb 7.3T 65G 7.3T 1% /data
/dev/sdm 7.3T 57G 7.3T 1% /data10
/dev/sdr 7.3T 62G 7.3T 1% /data15
/dev/sdg 7.3T 58G 7.3T 1% /data4
/dev/sdh 7.3T 59G 7.3T 1% /data5
/dev/sdq 7.3T 61G 7.3T 1% /data14
/dev/sdp 7.3T 57G 7.3T 1% /data13
/dev/sdf 7.3T 64G 7.3T 1% /data3
/dev/sdu 7.3T 60G 7.3T 1% /data18
/dev/sdj 7.3T 59G 7.3T 1% /data7
/dev/sdt 7.3T 68G 7.3T 1% /data17
/dev/sdl 7.3T 60G 7.3T 1% /data9
/dev/sdk 7.3T 58G 7.3T 1% /data8
/dev/sdd 7.3T 66G 7.3T 1% /data1
/dev/sde 7.3T 62G 7.3T 1% /data2
/dev/sdi 7.3T 79G 7.2T 2% /data6
/dev/sds 7.3T 65G 7.3T 1% /data16
/dev/mapper/rootvg-lv_var 173G 4.1G 169G 3% /var
/dev/mapper/rootvg-lv_opt 50G 4.5G 46G 9% /opt
tmpfs 19G 0 19G 0% /run/user/0
cm_processes 94G 8.8M 94G 1% /run/cloudera-scm-agent/process
tmpfs 19G 0 19G 0% /run/user/1001
$
其中，目录 /data19 相对其他数据目录过小，为统一存储目录空间大小保持一致，将其从 BE 的存储目录移除。具体操作流程，先停掉相应的 BE 服务，然后从 storage_root_path 中删掉对应盘的配置，然后重启 be 节点即可。但是做该操作之前一定要保证没有单副本的表，如果有的话，下掉该磁盘就很可能丢失数据，因为单副本的表的数据有可能就存储在要下掉的盘上。之后可以通过 show PROC '/statistic' 查看对应不健康的副本数。
查看对应不健康的副本数
$ mysql -h 192.168.37.10 -P9030 -uroot -p
Enter password:
Welcome to the MariaDB monitor. Commands end with ; or \g.
Your MySQL connection id is 7
Server version: 5.1.0 StarRocks version 2.2.1
Copyright (c) 2000, 2018, Oracle, MariaDB Corporation Ab and others.
Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.
MySQL [(none)]> show PROC '/statistic';
+-------+------------------------------+----------+--------------+----------+-----------+------------+--------------------+-----------------------+------------------+
| DbId | DbName | TableNum | PartitionNum | IndexNum | TabletNum | ReplicaNum | UnhealthyTabletNum | InconsistentTabletNum | CloningTabletNum |
+-------+------------------------------+----------+--------------+----------+-----------+------------+--------------------+-----------------------+------------------+
| 10002 | default_cluster:_statistics_ | 1 | 1 | 1 | 10 | 10 | 0 | 0 | 0 |
| 11001 | default_cluster:hive_test | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 10111 | default_cluster:dw | 1 | 61 | 61 | 610 | 1830 | 0 | 0 | 0 |
| Total | 3 | 2 | 62 | 62 | 620 | 1840 | 0 | 0 | 0 |
+-------+------------------------------+----------+--------------+----------+-----------+------------+--------------------+-----------------------+------------------+
4 rows in set (0.07 sec)
停止相应的 be 实例
$ cd be
$ ./bin/stop_be.sh
stop starrocks_be, and remove pid file.
$
再次查看对应不健康的副本数
MySQL [(none)]> show PROC '/statistic';
+-------+------------------------------+----------+--------------+----------+-----------+------------+--------------------+-----------------------+------------------+
| DbId | DbName | TableNum | PartitionNum | IndexNum | TabletNum | ReplicaNum | UnhealthyTabletNum | InconsistentTabletNum | CloningTabletNum |
+-------+------------------------------+----------+--------------+----------+-----------+------------+--------------------+-----------------------+------------------+
| 10002 | default_cluster:_statistics_ | 1 | 1 | 1 | 10 | 30 | 5 | 0 | 0 |
| 11001 | default_cluster:hive_test | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 10111 | default_cluster:dw | 1 | 61 | 61 | 610 | 1830 | 305 | 0 | 0 |
| Total | 3 | 2 | 62 | 62 | 620 | 1860 | 310 | 0 | 0 |
+-------+------------------------------+----------+--------------+----------+-----------+------------+--------------------+-----------------------+------------------+
4 rows in set (0.00 sec)
MySQL [(none)]>
修改相应 be 的配置
修改前：
$ cat conf/be.conf |grep -v '#'
sys_log_level = INFO
be_port = 9060
webserver_port = 18040
heartbeat_service_port = 9050
brpc_port = 8060
priority_networks = 192.168.37.16
storage_root_path = /data1/starrocks;/data2/starrocks;/data3/starrocks;/data4/starrocks;/data5/starrocks;/data6/starrocks;/data7/starrocks;/data8/starrocks;/data9/starrocks;/data10/starrocks;/data11/starrocks;/data12/starrocks;/data13/starrocks;/data14/starrocks;/data15/starrocks;/data16/starrocks;/data17/starrocks;/data18/starrocks;/data19/starrocks
default_rowset_type = beta
$
剔除掉目录 /data19/starrocks 后：
$ cat conf/be.conf |grep -v '#'
sys_log_level = INFO
be_port = 9060
webserver_port = 18040
heartbeat_service_port = 9050
brpc_port = 8060
priority_networks = 192.168.37.16
storage_root_path = /data1/starrocks;/data2/starrocks;/data3/starrocks;/data4/starrocks;/data5/starrocks;/data6/starrocks;/data7/starrocks;/data8/starrocks;/data9/starrocks;/data10/starrocks;/data11/starrocks;/data12/starrocks;/data13/starrocks;/data14/starrocks;/data15/starrocks;/data16/starrocks;/data17/starrocks;/data18/starrocks
default_rowset_type = beta
$
启动修改配置后的 be
$ ./bin/start_be.sh --daemon
再次查看对应不健康的副本数
MySQL [(none)]> show PROC '/statistic';
+-------+------------------------------+----------+--------------+----------+-----------+------------+--------------------+-----------------------+------------------+
| DbId | DbName | TableNum | PartitionNum | IndexNum | TabletNum | ReplicaNum | UnhealthyTabletNum | InconsistentTabletNum | CloningTabletNum |
+-------+------------------------------+----------+--------------+----------+-----------+------------+--------------------+-----------------------+------------------+
| 10002 | default_cluster:_statistics_ | 1 | 1 | 1 | 10 | 30 | 0 | 0 | 0 |
| 11001 | default_cluster:hive_test | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 10111 | default_cluster:dw | 1 | 61 | 61 | 610 | 1930 | 0 | 0 | 0 |
| Total | 3 | 2 | 62 | 62 | 620 | 1960 | 0 | 0 | 0 |
+-------+------------------------------+----------+--------------+----------+-----------+------------+--------------------+-----------------------+------------------+
4 rows in set (0.00 sec)
MySQL [(none)]>
等待 starrocks 副本机制自动修复不健康的副本数，即等待 UnhealthyTabletNum 列的值降至 0。
网友评论