标签:code down 就会 记录 点数据 协同 版本号 lin log
[admin@yrjk bin]$ ./zkCli.sh
[zk: localhost:2181(CONNECTED) 0]
[zk: localhost:2181(CONNECTED) 0] help ZooKeeper -server host:port cmd args stat path [watch] set path data [version] ls path [watch] delquota [-n|-b] path ls2 path [watch] setAcl path acl setquota -n|-b val path history redo cmdno printwatches on|off delete path [version] sync path listquota path rmr path get path [watch] create [-s] [-e] path data acl addauth scheme auth quit getAcl path close connect host:port [zk: localhost:2181(CONNECTED) 1]
[zk: localhost:2181(CONNECTED) 1] ls / # ls命令用于查看节点,类似于Linux中的查看目录 [zookeeper] [zk: localhost:2181(CONNECTED) 2] ls2 / # ls2命令用于查看节点以及该节点状态的详细信息 [zookeeper] cZxid = 0x0 ctime = Thu Jan 01 08:00:00 CST 1970 mZxid = 0x0 mtime = Thu Jan 01 08:00:00 CST 1970 pZxid = 0x0 cversion = -1 dataVersion = 0 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 0 numChildren = 1 [zk: localhost:2181(CONNECTED) 3]
节点状态的属性描述如下:
[zk: localhost:2181(CONNECTED) 3] create /name Richard Created /name [zk: localhost:2181(CONNECTED) 4] stat /name # stat命令用于查看节点状态的详细信息 cZxid = 0x6 ctime = Sun Dec 15 13:25:02 CST 2019 mZxid = 0x6 mtime = Sun Dec 15 13:25:02 CST 2019 pZxid = 0x6 cversion = 0 dataVersion = 0 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 7 numChildren = 0 [zk: localhost:2181(CONNECTED) 5] get /name # get命令用于查看节点的数据以及节点状态的详细信息 Richard cZxid = 0x6 ctime = Sun Dec 15 13:25:02 CST 2019 mZxid = 0x6 mtime = Sun Dec 15 13:25:02 CST 2019 pZxid = 0x6 cversion = 0 dataVersion = 0 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 7 numChildren = 0 [zk: localhost:2181(CONNECTED) 6]
znode一共有4种类型:持久的(persistent)、临时的(ephemeral)、持久有序的(persistent_sequential)和临时有序的(ephemeral_sequential)。
在节点创建之前,需要了解一下zk中session的特性:
[zk: localhost:2181(CONNECTED) 2] create /name Richard Created /name [zk: localhost:2181(CONNECTED) 3] ls / [name, zookeeper] [zk: localhost:2181(CONNECTED) 4] get /name Richard cZxid = 0xe ctime = Sun Dec 15 13:53:22 CST 2019 mZxid = 0xe mtime = Sun Dec 15 13:53:22 CST 2019 pZxid = 0xe cversion = 0 dataVersion = 0 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 7 numChildren = 0 [zk: localhost:2181(CONNECTED) 5]
这种创建方式创建出来的节点是持久化的,也就是持久节点(PERSISTENT)。所谓持久节点,是指在节点创建后,就一直存在,直到有删除操作来主动清除这个节点——不会因为创建该节点的客户端会话失效而消失。除了持久节点之外,我们还可以创建临时节点(EPHEMERAL),那么我们来看看如何创建临时节点:
[zk: localhost:2181(CONNECTED) 5] create -e /name/tmp Rachel Created /name/tmp # -e 指定创建临时节点 [zk: localhost:2181(CONNECTED) 6] get /name Richard cZxid = 0xe ctime = Sun Dec 15 13:53:22 CST 2019 mZxid = 0xe mtime = Sun Dec 15 13:53:22 CST 2019 pZxid = 0xf cversion = 1 #由于在/name下创建了一个子节点,所以 cversion 的值就会进行累加 dataVersion = 0 aclVersion = 0 ephemeralOwner = 0x0 #表示持久节点 dataLength = 7 numChildren = 1 [zk: localhost:2181(CONNECTED) 7] get /name/tmp Rachel cZxid = 0xf ctime = Sun Dec 15 13:56:09 CST 2019 mZxid = 0xf mtime = Sun Dec 15 13:56:09 CST 2019 pZxid = 0xf cversion = 0 dataVersion = 0 aclVersion = 0 ephemeralOwner = 0x100000be9f80004 #表示临时节点 dataLength = 6 numChildren = 0 [zk: localhost:2181(CONNECTED) 8]
和持久节点不同的是,临时节点的生命周期和客户端会话绑定。也就是说,如果客户端会话失效,那么这个节点就会自动被清除掉。注意,这里提到的是会话失效,而非连接断开,当然连接断开也会导致会话失效,但是并不是主要原因。另外,在临时节点下面不能创建子节点。我们上面提到了,心跳结束,session就会过期,而session过期,则临时节点znode就会被抛弃。那么我们来断开与服务端的连接,看看临时节点是否会被清除:
[zk: localhost:2181(CONNECTED) 8] ls /name [tmp] [zk: localhost:2181(CONNECTED) 9] quit Quitting... 2019-12-15 14:00:44,326 [myid:] - INFO [main:ZooKeeper@687] - Session: 0x100000be9f80004 closed 2019-12-15 14:00:44,328 [myid:] - INFO [main-EventThread:ClientCnxn$EventThread@520] - EventThread shut down for session: 0x100000be9f80004 [admin@yrjk bin]$ ./zkCli.sh [zk: localhost:2181(CONNECTED) 0] ls /name #重连之后,临时节点被清除 [] [zk: localhost:2181(CONNECTED) 1]
以上我们演示了持久节点和临时节点的创建,下面我们来看一下持久顺序节点(PERSISTENT_SEQUENTIAL)的创建:
[zk: localhost:2181(CONNECTED) 3] create -s /name/seq num #-s指定创建持久顺序节点 Created /name/seq0000000001 # 会自动为给定节点名加上一个数字后缀 [zk: localhost:2181(CONNECTED) 4] ls /name [seq0000000001] [zk: localhost:2181(CONNECTED) 5] create -s /name/seq num Created /name/seq0000000002 # 再次创建节点数字就会递增 [zk: localhost:2181(CONNECTED) 6] ls /name [seq0000000001, seq0000000002] [zk: localhost:2181(CONNECTED) 7] create -s /name/otherseq num Created /name/otherseq0000000003 # 创建前缀不同的节点,数字也会递增 [zk: localhost:2181(CONNECTED) 8] ls /name [otherseq0000000003, seq0000000001, seq0000000002] [zk: localhost:2181(CONNECTED) 9]
这类节点的基本特性和持久节点类型是一致的。额外的特性是,在ZK中,每个父节点会为他的第一级子节点维护一份时序,会记录每个子节点创建的先后顺序。基于这个特性,在创建子节点的时候,可以设置这个属性,那么在创建节点过程中,ZK会自动为给定节点名加上一个数字后缀,作为新的节点名。这个数字后缀的范围是整型的最大值。
当 -s 与 -e 选项同时使用就是创建临时顺序节点(EPHEMERAL_SEQUENTIAL) ,此节点是属于临时节点,不过带有顺序,和临时节点一样,当session过期节点就会消失,而客户端会话连接结束也会导致session过期,所以同样的该节点也会消失,这种类型的节点一般用于实现分布式锁。以下演示一下临时顺序节点的创建方式:
[zk: localhost:2181(CONNECTED) 9] create /address Beijing #创建一个持久节点 Created /address [zk: localhost:2181(CONNECTED) 10] create -s -e /address/road Changan #在持久节点下,创建临时顺序节点 Created /address/road0000000000 [zk: localhost:2181(CONNECTED) 11] create -s -e /address/road Changan Created /address/road0000000001 [zk: localhost:2181(CONNECTED) 12] create -s -e /address/street Changan Created /address/street0000000002 [zk: localhost:2181(CONNECTED) 13] ls /address [road0000000000, road0000000001, street0000000002] [zk: localhost:2181(CONNECTED) 14] quit Quitting... 2019-12-15 14:15:09,721 [myid:] - INFO [main:ZooKeeper@687] - Session: 0x100000be9f80005 closed 2019-12-15 14:15:09,723 [myid:] - INFO [main-EventThread:ClientCnxn$EventThread@520] - EventThread shut down for session: 0x100000be9f80005 [admin@yrjk bin]$ ./zkCli.sh [zk: localhost:2181(CONNECTED) 0] ls /address []
使用set命令对某个节点进行修改:
[zk: localhost:2181(CONNECTED) 6] get /name Richard cZxid = 0xe ctime = Sun Dec 15 13:53:22 CST 2019 mZxid = 0xe mtime = Sun Dec 15 13:53:22 CST 2019 pZxid = 0x14 cversion = 5 dataVersion = 0 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 7 numChildren = 3 [zk: localhost:2181(CONNECTED) 7] set /name David #修改/name节点的数据 cZxid = 0xe ctime = Sun Dec 15 13:53:22 CST 2019 mZxid = 0x1b mtime = Sun Dec 15 14:19:21 CST 2019 pZxid = 0x14 cversion = 5 dataVersion = 1 # 此时数据版本就会递增为1 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 5 numChildren = 3 [zk: localhost:2181(CONNECTED) 8] get /name David cZxid = 0xe ctime = Sun Dec 15 13:53:22 CST 2019 mZxid = 0x1b mtime = Sun Dec 15 14:19:21 CST 2019 pZxid = 0x14 cversion = 5 dataVersion = 1 # 数据版本递增为1 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 5 numChildren = 3 [zk: localhost:2181(CONNECTED) 9]
我们可以利用数据版本 dataVersion 来达到一个乐观锁的效果,所以每次我们修改节点数据的时候,应该加上这个 dataVersion 的值去进行修改,以免在并发的时候导致数据不一致:
[zk: localhost:2181(CONNECTED) 9] set /name Sam 1 #指定版本去修改 cZxid = 0xe ctime = Sun Dec 15 13:53:22 CST 2019 mZxid = 0x1c mtime = Sun Dec 15 14:22:33 CST 2019 pZxid = 0x14 cversion = 5 dataVersion = 2 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 3 numChildren = 3 [zk: localhost:2181(CONNECTED) 10] set /name Sam1 1 # 如果再次使用版本1去修改数据,就会报错 version No is not valid : /name [zk: localhost:2181(CONNECTED) 11]
使用delete命令可以对某个节点进行删除:
[zk: localhost:2181(CONNECTED) 12] ls /name [otherseq0000000003, seq0000000001, seq0000000002] [zk: localhost:2181(CONNECTED) 13] delete /name/seq0000000001 #删除节点 [zk: localhost:2181(CONNECTED) 14] ls /name [otherseq0000000003, seq0000000002] [zk: localhost:2181(CONNECTED) 15] delete /name/seq0000000002 0 #指定版本号删除节点 [zk: localhost:2181(CONNECTED) 16] ls /name [otherseq0000000003]
使用rmr命令可以对节点进行递归删除:
[zk: localhost:2181(CONNECTED) 16] ls / [name, address, zookeeper] [zk: localhost:2181(CONNECTED) 17] rmr /name #递归删除节点 [zk: localhost:2181(CONNECTED) 18] ls / [address, zookeeper] [zk: localhost:2181(CONNECTED) 19]
watcher是zk中比较重要的特性,定义如下:
watcher命令行学习:
watcher事件类型-父节点:
创建父节点触发 NodeCreated 事件,示例:
[zk: localhost:2181(CONNECTED) 19] stat /testWatch watch # 在节点创建之前,我们可以通过 stat 命令去设置watcher Node does not exist: /testWatch [zk: localhost:2181(CONNECTED) 20] create /testWatch test-data # 创建父节点 WATCHER:: WatchedEvent state:SyncConnected type:NodeCreated path:/testWatch # 触发 NodeCreated 事件 Created /testWatch [zk: localhost:2181(CONNECTED) 21]
修改父节点数据触发 NodeDataChanged 事件,示例:
[zk: localhost:2181(CONNECTED) 21] get /testWatch watch # 因为zk事件是一次性的,所以我们还需要通过 get 命令设置 watcher test-data cZxid = 0x19 ctime = Sun Apr 22 23:37:08 CST 2018 mZxid = 0x19 mtime = Sun Apr 22 23:37:08 CST 2018 pZxid = 0x19 cversion = 0 dataVersion = 0 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 9 numChildren = 0 [zk: localhost:2181(CONNECTED) 22] set /testWatch new-data # 修改父节点数据 WATCHER:: WatchedEvent state:SyncConnected type:NodeDataChanged path:/testWatch # 触发 NodeDataChanged 事件 cZxid = 0x19 ctime = Sun Apr 22 23:37:08 CST 2018 mZxid = 0x1a mtime = Sun Apr 22 23:40:32 CST 2018 pZxid = 0x19 cversion = 0 dataVersion = 1 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 8 numChildren = 0 [zk: localhost:2181(CONNECTED) 23]
删除父节点触发 NodeDeleted 事件,示例:
[zk: localhost:2181(CONNECTED) 23] ls /testWatch watch # 通过 ls 命令来设置 watcher [] [zk: localhost:2181(CONNECTED) 24] delete /testWatch # 删除父节点 WATCHER:: WatchedEvent state:SyncConnected type:NodeDeleted path:/testWatch # 触发 NodeDeleted 事件 [zk: localhost:2181(CONNECTED) 25]
watcher事件类型-子节点:
使用 ls 命令为父节点设置watcher,创建子节点时就会触发 NodeChildrenChanged 事件,示例:
[zk: localhost:2181(CONNECTED) 29] create /testWatch test-data # 创建父节点 Created /testWatch [zk: localhost:2181(CONNECTED) 30] ls /testWatch watch # 使用 ls 命令为父节点设置watcher [] [zk: localhost:2181(CONNECTED) 31] create /testWatch/testChildren children-data # 创建子节点 WATCHER:: WatchedEvent state:SyncConnected type:NodeChildrenChanged path:/testWatch # 触发 NodeChildrenChanged 事件 Created /testWatch/testChildren [zk: localhost:2181(CONNECTED) 32]
使用 ls 命令为父节点设置watcher,删除子节点时也会触发 NodeChildrenChanged 事件,示例:
[zk: localhost:2181(CONNECTED) 32] ls /testWatch watch # 使用 ls 命令为父节点设置watcher [testChildren] [zk: localhost:2181(CONNECTED) 33] delete /testWatch/testChildren # 删除子节点 WATCHER:: WatchedEvent state:SyncConnected type:NodeChildrenChanged path:/testWatch # 触发 NodeChildrenChanged 事件 [zk: localhost:2181(CONNECTED) 34]
简单说明一下为什么创建和删除子节点都是触发 NodeChildrenChanged 事件,这是因为子节点都是挂在父节点之下,而我们是给父节点设置的 watcher,不是给子节点设置 watcher ,不管子节点是删除还是创建,都是一个改变的过程,所以都是触发同一个事件。
使用 ls 命令为父节点设置watcher,修改子节点数据时不会触发任何事件,示例:
[zk: localhost:2181(CONNECTED) 35] create /testWatch/testChildren children-data # 创建子节点 [zk: localhost:2181(CONNECTED) 36] ls /testWatch watch # 使用 ls 命令为父节点设置watcher [testChildren] [zk: localhost:2181(CONNECTED) 37] set /testWatch/testChildren new-children-data # 修改子节点数据时不会触发任何事件 cZxid = 0x1f ctime = Sun Apr 22 23:58:44 CST 2018 mZxid = 0x20 mtime = Sun Apr 22 23:59:24 CST 2018 pZxid = 0x1f cversion = 0 dataVersion = 1 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 17 numChildren = 0 [zk: localhost:2181(CONNECTED) 38]
不会触发事件是因为这个watcher是设置在父节点上的,所以修改子节点数据时不会触发父节点所设置的watcher事件。
使用 get 命令为子节点设置watcher,修改子节点数据时会触发 NodeDataChanged 事件,示例:
[zk: localhost:2181(CONNECTED) 40] get /testWatch/testChildren watch # 使用 get 命令为子节点设置watcher new-children-data cZxid = 0x1f ctime = Sun Apr 22 23:58:44 CST 2018 mZxid = 0x21 mtime = Mon Apr 23 00:01:41 CST 2018 pZxid = 0x1f cversion = 0 dataVersion = 2 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 17 numChildren = 0 [zk: localhost:2181(CONNECTED) 41] set /testWatch/testChildren new-children-data2 # 修改子节点数据 WATCHER:: WatchedEvent state:SyncConnected type:NodeDataChanged path:/testWatch/testChildren # 触发 NodeDataChanged 事件 cZxid = 0x1f ctime = Sun Apr 22 23:58:44 CST 2018 mZxid = 0x22 mtime = Mon Apr 23 00:02:11 CST 2018 pZxid = 0x1f cversion = 0 dataVersion = 3 aclVersion = 0 ephemeralOwner = 0x0 dataLength = 18 numChildren = 0 [zk: localhost:2181(CONNECTED) 42]
参考:ZooKeeper:分布式过程协同技术详解
https://blog.51cto.com/zero01/2106494
标签:code down 就会 记录 点数据 协同 版本号 lin log
原文地址:https://www.cnblogs.com/zjfjava/p/11909389.html