001/** 002 * Licensed to the Apache Software Foundation (ASF) under one 003 * or more contributor license agreements. See the NOTICE file 004 * distributed with this work for additional information 005 * regarding copyright ownership. The ASF licenses this file 006 * to you under the Apache License, Version 2.0 (the 007 * "License"); you may not use this file except in compliance 008 * with the License. You may obtain a copy of the License at 009 * 010 * http://www.apache.org/licenses/LICENSE-2.0 011 * 012 * Unless required by applicable law or agreed to in writing, software 013 * distributed under the License is distributed on an "AS IS" BASIS, 014 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 015 * See the License for the specific language governing permissions and 016 * limitations under the License. 017 */ 018 019package org.apache.hadoop.hbase.chaos.actions; 020 021import java.io.IOException; 022 023import org.apache.hadoop.hbase.ServerName; 024import org.apache.hadoop.hbase.util.Threads; 025import org.slf4j.Logger; 026import org.slf4j.LoggerFactory; 027 028/** 029* Base class for restarting HBaseServer's 030*/ 031public class RestartActionBaseAction extends Action { 032 private static final Logger LOG = 033 LoggerFactory.getLogger(RestartActionBaseAction.class); 034 long sleepTime; // how long should we sleep 035 036 public RestartActionBaseAction(long sleepTime) { 037 this.sleepTime = sleepTime; 038 } 039 040 void sleep(long sleepTime) { 041 LOG.info("Sleeping for:" + sleepTime); 042 Threads.sleep(sleepTime); 043 } 044 045 void restartMaster(ServerName server, long sleepTime) throws IOException { 046 sleepTime = Math.max(sleepTime, 1000); 047 // Don't try the kill if we're stopping 048 if (context.isStopping()) { 049 return; 050 } 051 052 LOG.info("Killing master: {}", server); 053 killMaster(server); 054 sleep(sleepTime); 055 LOG.info("Starting master: {}", server); 056 startMaster(server); 057 } 058 059 /** 060 * Stop and then restart the region server instead of killing it. 061 * @param server hostname to restart the regionserver on 062 * @param sleepTime number of milliseconds between stop and restart 063 * @throws IOException if something goes wrong 064 */ 065 void gracefulRestartRs(ServerName server, long sleepTime) throws IOException { 066 sleepTime = Math.max(sleepTime, 1000); 067 // Don't try the stop if we're stopping already 068 if (context.isStopping()) { 069 return; 070 } 071 LOG.info("Stopping region server: {}", server); 072 stopRs(server); 073 sleep(sleepTime); 074 LOG.info("Starting region server: {}", server); 075 startRs(server); 076 } 077 078 void restartRs(ServerName server, long sleepTime) throws IOException { 079 sleepTime = Math.max(sleepTime, 1000); 080 // Don't try the kill if we're stopping 081 if (context.isStopping()) { 082 return; 083 } 084 LOG.info("Killing region server: {}", server); 085 killRs(server); 086 sleep(sleepTime); 087 LOG.info("Starting region server: {}", server); 088 startRs(server); 089 } 090 091 void restartZKNode(ServerName server, long sleepTime) throws IOException { 092 sleepTime = Math.max(sleepTime, 1000); 093 // Don't try the kill if we're stopping 094 if (context.isStopping()) { 095 return; 096 } 097 LOG.info("Killing zookeeper node: {}", server); 098 killZKNode(server); 099 sleep(sleepTime); 100 LOG.info("Starting zookeeper node: {}", server); 101 startZKNode(server); 102 } 103 104 void restartDataNode(ServerName server, long sleepTime) throws IOException { 105 sleepTime = Math.max(sleepTime, 1000); 106 // Don't try the kill if we're stopping 107 if (context.isStopping()) { 108 return; 109 } 110 LOG.info("Killing data node: {}", server); 111 killDataNode(server); 112 sleep(sleepTime); 113 LOG.info("Starting data node: {}", server); 114 startDataNode(server); 115 } 116 117 void restartNameNode(ServerName server, long sleepTime) throws IOException { 118 sleepTime = Math.max(sleepTime, 1000); 119 // Don't try the kill if we're stopping 120 if (context.isStopping()) { 121 return; 122 } 123 LOG.info("Killing name node: {}", server); 124 killNameNode(server); 125 sleep(sleepTime); 126 LOG.info("Starting name node: {}", server); 127 startNameNode(server); 128 } 129 130}