使用./bin/graceful_stop.sh had1停止一个hbase regionserver失败

日记如下:
2016-03-02 19:33:26,580 INFO zookeeper.ZooKeeper: Client environment:user.dir=/home/chris/hbase-0.98.7-hadoop2/bin
2016-03-02 19:33:26,581 INFO zookeeper.ZooKeeper: Initiating client connection, connectString=10.1.41.94:2181,10.1.55.56:2181,10.1.59.225:2181 sessionTimeout=90000 watcher=hconnection-0x661261ba, quorum=10.1.41.94:2181,10.1.55.56:2181,10.1.59.225:2181, baseZNode=/hbase
2016-03-02 19:33:26,603 INFO zookeeper.ClientCnxn: Opening socket connection to server 10.1.59.225/10.1.59.225:2181. Will not attempt to authenticate using SASL (unknown error)
2016-03-02 19:33:26,610 INFO zookeeper.ClientCnxn: Socket connection established to 10.1.59.225/10.1.59.225:2181, initiating session
2016-03-02 19:33:26,617 INFO zookeeper.ClientCnxn: Session establishment complete on server 10.1.59.225/10.1.59.225:2181, sessionid = 0x3500747e7e6841d, negotiated timeout = 40000
RuntimeError: Server had1:60020 not online
stripServer at /home/chris/hbase-0.98.7-hadoop2/bin/region_mover.rb:225
unloadRegions at /home/chris/hbase-0.98.7-hadoop2/bin/region_mover.rb:336
(root) at /home/chris/hbase-0.98.7-hadoop2/bin/region_mover.rb:515
2015-11-27T06:26:00 Unloaded had1 region(s)
2015-11-27T06:26:00 Stopping regionserver
no regionserver to stop because no pid file /tmp/hbase-root-regionserver.pid
2015-11-27T06:26:00 Restoring balancer state to端口测试连接也是没有问题的:
root@kyo:/home/chris/hadoop-2.2.0/etc/hadoop# nc -v -z had1 60020
Connection to had1 60020 port succeeded!
已邀请:

采菊篱下 - 无分享不快乐 助力技术开源分享 运维一枚

赞同来自:

判断你是在HMaster上执行的./bin/graceful_stop.sh had1的脚本,判断原因如下:
1、日记中报:RuntimeError: Server had1:60020 not online,但是你说的你探测端口是正常的
2、no regionserver to stop because no pid file /tmp/hbase-root-regionserver.pid
没有找到pid文件,被误删除了有可能,但是如果没有清理的脚本或者程序,消失的可能性也很小。
所以正确的做法是,你应该到hbase regionserver节点had1上面去执行脚本:
./bin/graceful_stop.sh had1

要回复问题请先登录注册