运维 CLI
Contents
运维 CLI#
NS Client#
连接到ns client 需要指定zk_cluster、zk_root_path和role。其中zk_cluster是zk地址,zk_root_path是集群在zk的根路径,role是启动的角色需指定为ns_client
$ ./bin/openmldb --zk_cluster=172.27.2.52:12200 --zk_root_path=/onebox --role=ns_client
use#
use命令可以切换到某个database下
> use demodb
showtable#
showtable可以查看所有表,也可以指定看某一个表
命令格式:showtable [table_name]
172.24.4.55:6531 demo_db> showtable
name tid pid endpoint role is_alive offset record_cnt memused diskused
------------------------------------------------------------------------------------------------------
aaa 22 0 172.24.4.55:9971 leader yes 0 0 0.000 8.646 K
auto_VCeOIIKA 25 0 172.24.4.55:9971 leader yes 8 4 498.000 9.128 K
t1 21 0 172.24.4.55:9971 leader yes 1 1 301.000 9.353 K
172.24.4.55:6531 demo_db> showtable auto_VCeOIIKA
name tid pid endpoint role is_alive offset record_cnt memused diskused
------------------------------------------------------------------------------------------------------
auto_VCeOIIKA 25 0 172.24.4.55:9971 leader yes 8 4 498.000 9.128 K
showtablet#
查看tablet信息(如果使用的serverName和自动获取本地ip功能,则endpoint为serverName, real_endpoint为“-”)
> showtablet
endpoint real_endpoint state age
--------------------------------------------------------------
6534708411798331392 172.17.0.12:9531 kTabletHealthy 1d
6534708415948603392 172.17.0.13:9532 kTabletHealthy 1d
6534708420092481536 172.17.0.14:9533 kTabletHealthy 14h
addreplica#
添加副本
命令格式: addreplica table_name pid_group endpoint
table_name 表名
pid_group 分片id集合. 可以有以下几种情况
单个分片
多个分片, 分片间用逗号分割. 如1,3,5
分片区间, 区间为闭区间. 如1-5表示分片1,2,3,4,5
endpoint 要添加为副本的节点endpoint
> showtable test1
name tid pid endpoint role ttl is_alive compress_type offset record_cnt memused
------------------------------------------------------------------------------------------------------------
test1 13 0 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
test1 13 1 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
test1 13 2 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
test1 13 3 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
> addreplica test1 0 172.27.128.33:8541
AddReplica ok
> showtable test1
name tid pid endpoint role ttl is_alive compress_type offset record_cnt memused
------------------------------------------------------------------------------------------------------------
test1 13 0 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
test1 13 0 172.27.128.33:8541 follower 0min yes kNoCompress 0 0 0.000
test1 13 1 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
test1 13 2 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
test1 13 3 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
> addreplica test1 1,2,3 172.27.128.33:8541
AddReplica ok
> addreplica test1 1-3 172.27.128.33:8541
AddReplica ok
delreplica#
删除副本
命令格式: delreplica table_name pid_group endpoint
table_name 表名
pid_group 分片id集合. 可以有以下几种情况
单个分片
多个分片, 分片间用逗号分割. 如1,3,5
分片区间, 区间为闭区间. 如1-5表示分片1,2,3,4,5
endpoint 要删除副本的endpoint
> showtable test1
name tid pid endpoint role ttl is_alive compress_type offset record_cnt memused
------------------------------------------------------------------------------------------------------------
test1 13 0 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
test1 13 0 172.27.128.33:8541 follower 0min yes kNoCompress 0 0 0.000
test1 13 1 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
test1 13 1 172.27.128.33:8541 follower 0min yes kNoCompress 0 0 0.000
test1 13 2 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
test1 13 2 172.27.128.33:8541 follower 0min yes kNoCompress 0 0 0.000
test1 13 3 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
test1 13 3 172.27.128.33:8541 follower 0min yes kNoCompress 0 0 0.000
> delreplica test1 0 172.27.128.33:8541
DelReplica ok
> showtable test1
name tid pid endpoint role ttl is_alive compress_type offset record_cnt memused
------------------------------------------------------------------------------------------------------------
test1 13 0 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
test1 13 1 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
test1 13 1 172.27.128.33:8541 follower 0min yes kNoCompress 0 0 0.000
test1 13 2 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
test1 13 2 172.27.128.33:8541 follower 0min yes kNoCompress 0 0 0.000
test1 13 3 172.27.128.31:8541 leader 0min yes kNoCompress 0 0 0.000
test1 13 3 172.27.128.33:8541 follower 0min yes kNoCompress 0 0 0.000
> delreplica test1 1,2,3 172.27.128.33:8541
DelReplica ok
> delreplica test1 1-3 172.27.128.33:8541
DelReplica ok
migrate#
副本迁移
命令格式: migrate src_endpoint table_name pid_group des_endpoint
src_endpoint 需要迁出的节点
table_name 表名
pid_group 分片id集合. 可以有以下几种情况
单个分片
多个分片, 分片间用逗号分割. 如1,3,5
分片区间, 区间为闭区间. 如1-5表示分片1,2,3,4,5
des_endpoint 迁移的目的节点
> migrate 172.27.2.52:9991 table1 1 172.27.2.52:9992
partition migrate ok
> migrate 172.27.2.52:9991 table1 1-5 172.27.2.52:9992
partition migrate ok
> migrate 172.27.2.52:9991 table1 1,2,3 172.27.2.52:9992
partition migrate ok
confget#
获取配置信息,目前只支持auto_failover
命令格式: confget [conf_name]
conf_name 配置项名字,可选的
> confget
key value
-----------------------------
auto_failover false
> confget auto_failover
key value
------------------------
auto_failover false
confset#
修改配置信息,目前只支持auto_failover
命令格式: confset conf_name value
conf_name 配置项名字
value 配置项设置的值
> confset auto_failover true
set auto_failover ok
offlineendpoint#
下线节点。此命令是异步的返回成功后可通过showopstatus查看运行状态
命令格式: offlineendpoint endpoint [concurrency]
endpoint是发生故障节点的endpoint。该命令会对该节点下所有分片执行如下操作:
如果是主, 执行重新选主
如果是从, 找到主节点然后从主节点中删除当前endpoint副本
修改is_alive状态为no
concurrency 控制任务执行的并发数. 此配置是可选的, 默认为2(name_server_task_concurrency配置可配), 最大值为name_server_task_max_concurrency配置的值
> offlineendpoint 172.27.128.32:8541
offline endpoint ok
>showtable
name tid pid endpoint role ttl is_alive compress_type offset record_cnt memused
----------------------------------------------------------------------------------------------------------------------
flow 4 0 172.27.128.32:8541 leader 0min no kNoCompress 0 0 0.000
flow 4 0 172.27.128.33:8541 follower 0min yes kNoCompress 0 0 0.000
flow 4 0 172.27.128.31:8541 follower 0min yes kNoCompress 0 0 0.000
flow 4 1 172.27.128.33:8541 leader 0min yes kNoCompress 0 0 0.000
flow 4 1 172.27.128.31:8541 follower 0min yes kNoCompress 0 0 0.000
flow 4 1 172.27.128.32:8541 follower 0min no kNoCompress 0 0 0.000
该命令执行成功后所有分片都会有yes状态的leader
recoverendpoint#
恢复节点数据。此命令是异步的返回成功后可通过showopstatus查看运行状态
命令格式: recoverendpoint endpoint [need_restore] [concurrency]
endpoint是要恢复节点的endpoint
need_restore 表拓扑是否要恢复到最初的状态, 此配置是可选的, 默认为false. 如果设置为true, 一个分片在该节点下为leader, 执行完recoverendpoint恢复数据后依然是leader
concurrency 控制任务执行的并发数. 此配置是可选的, 默认为2(name_server_task_concurrency配置可配), 最大值为name_server_task_max_concurrency配置的值
> recoverendpoint 172.27.128.32:8541
recover endpoint ok
> recoverendpoint 172.27.128.32:8541 true
recover endpoint ok
> recoverendpoint 172.27.128.32:8541 true 3
recover endpoint ok
注:
执行此命令前确保节点已经上线(showtablet命令查看)
changeleader#
对某个指定的分片执行主从切换。此命令是异步的返回成功后可通过showopstatus查看运行状态
命令格式: changeleader table_name pid [candidate_leader]
table_name 表名
pid 分片id
candidate_leader 候选leader. 该参数是可选的。如果不加该参数则要求分片中不存在alive是yes的leader,如果设置成auto即使其他节点alive状态是yes也能切换
> changeleader flow 0
change leader ok
> changeleader flow 0 172.27.128.33:8541
change leader ok
> changeleader flow 0 auto
change leader ok
recovertable#
恢复某个分片数据。此命令是异步的返回成功后可通过showopstatus查看运行状态
命令格式: recovertable table_name pid endpoint
table_name 表名
pid 分片id
endpoint 要恢复分片所在的节点endpoint
> recovertable flow 1 172.27.128.31:8541
recover table ok
cancelop#
取消一个正在执行或者待执行的操作. 取消后任务就状态变成kCanceled
命令格式: cancelop op_id
op_id 需要取消的操作id
> cancelop 5
Cancel op ok!
showopstatus#
显示操作执行信息
命令格式: showopstatus [table_name pid]
table_name 表名
pid 分片id
> showopstatus
op_id op_type name pid status start_time execute_time end_time cur_task
------------------------------------------------------------------------------------------------------------
51 kMigrateOP flow 0 kDone 20180824163316 12s 20180824163328 -
52 kRecoverTableOP flow 0 kDone 20180824195252 1s 20180824195253 -
53 kRecoverTableOP flow 1 kDone 20180824195252 1s 20180824195253 -
54 kUpdateTableAliveOP flow 0 kDone 20180824195838 2s 20180824195840 -
55 kChangeLeaderOP flow 0 kDone 20180824200135 0s 20180824200135 -
56 kOfflineReplicaOP flow 1 kDone 20180824200135 1s 20180824200136 -
57 kReAddReplicaOP flow 0 kDone 20180827114331 12s 20180827114343 -
58 kAddReplicaOP test1 0 kDone 20180827205907 8s 20180827205915 -
59 kDelReplicaOP test1 0 kDone 20180827210248 4s 20180827210252 -
> showopstatus flow
op_id op_type name pid status start_time execute_time end_time cur_task
-----------------------------------------------------------------------------------------------------------
51 kMigrateOP flow 0 kDone 20180824163316 12s 20180824163328 -
52 kRecoverTableOP flow 0 kDone 20180824195252 1s 20180824195253 -
53 kRecoverTableOP flow 1 kDone 20180824195252 1s 20180824195253 -
54 kUpdateTableAliveOP flow 0 kDone 20180824195838 2s 20180824195840 -
55 kChangeLeaderOP flow 0 kDone 20180824200135 0s 20180824200135 -
56 kOfflineReplicaOP flow 1 kDone 20180824200135 1s 20180824200136 -
57 kUpdateTableAliveOP flow 0 kDone 20180824200212 0s 20180824200212 -
>showopstatus flow 1
op_id op_type name pid status start_time execute_time end_time cur_task
-----------------------------------------------------------------------------------------------------------
53 kRecoverTableOP flow 1 kDone 20180824195252 1s 20180824195253 -
56 kOfflineReplicaOP flow 1 kDone 20180824200135 1s 20180824200136 -
updatetablealive#
修改分片alive状态
命令格式: updatetablealive table_name pid endpoint is_alive
table_name 表名
pid 分片id, 如果要修改某个表的所有分片将pid指定为*
endpoint 节点endpoint
is_alive 节点状态, 只能填yes或者no
> updatetablealive test * 172.27.128.31:8541 no
update ok
> updatetablealive test 1 172.27.128.31:8542 no
update ok
注: 此命令不可当做故障恢复使用. 一般用于切流量,操作方式为将某个节点表的alive状态改为no读请求就不会落在该节点上
showns#
显示nameserver节点及其角色(如果使用的serverName和自动获取本地ip功能,则endpoint为serverName, real_endpoint为“-”)
命令格式: showns
>showns
endpoint real_endpoint role
-------------------------------------------
172.24.4.55:6531 - leader
exit#
退出客户端
> exit
bye
quit#
退出客户端
> quit
bye
Tablet Client#
连接到tablet client需要指定endpoint和role. 其中endpoint是需要连接tablet的endpoint,role是启动的角色需指定为client
$ ./openmldb --endpoint=172.27.2.52:9520 --role=client
loadtable#
1、加载已有表
命令格式: loadtable table_name tid pid ttl segment_cnt
table_name 表名
tid 指定table的id
pid 指定table的分片id
ttl 指定ttl
segment_cnt 指定segment_cnt, 一般设置为8
> loadtable table1 1 0 144000 8
已有表如果在内存中则loadtable会失败
changerole#
改变表的leader角色
命令格式: changerole tid pid role [term]
tid 指定table的id
pid 指定table的分片id
role 要修改的角色, 取值为[leader, follower]
term 设定term, 该项是可选的, 默认为0
> changerole 1 0 follower
ChangeRole ok
> changerole 1 0 leader
ChangeRole ok
> changerole 1 0 leader 1002
ChangeRole ok
gettablestatus#
获取表信息
命令格式: gettablestatus [tid pid]
tid 指定table的id
pid 指定table的分片id
> gettablestatus
tid pid offset mode state enable_expire ttl ttl_offset memused compress_type
----------------------------------------------------------------------------------------------------------------------------
1 0 4 kTableLeader kSnapshotPaused true 144000min 0s 1.313 K kNoCompress
2 0 4 kTableLeader kTableNormal false 0min 0s 689.000 kNoCompress
> gettablestatus 2 0
tid pid offset mode state enable_expire ttl ttl_offset memused compress_type
---------------------------------------------------------------------------------------------------------
2 0 4 kTableLeader kTableNormal false 0min 0s 689.000 kNoCompress
getfollower#
查看从节点信息
命令格式: getfollower tid pid
tid 指定table的id
pid 指定table的分片id
> getfollower 4 1
# tid pid leader_offset follower offset
-----------------------------------------------------------
0 4 1 5923724 172.27.128.31:8541 5920714
1 4 1 5923724 172.27.128.32:8541 5921707
exit#
退出客户端
> exit
bye
quit#
退出客户端
> quit
bye