def localStop(self):
nodeId = dbClusterInfo.getNodeIdByName(self.dbNodeInfo.name, self.clusterConfig)
cmd = ClusterCommand.getStopCmd(self.user, nodeId, "i")
(status, output) = commands.getstatusoutput(cmd)
if (status != 0):
self.logger.log(output)
self.logger.logExit("Stop node failed!")
self.logger.log("Stop node success.")
self.logger.closeLog()
def __switchToPrimary(self, datadir):
"""
Switch local instance to be primary
"""
cmd = ClusterCommand.getSwitchOverCmd(self.user, self.dbNodeInfo.id, datadir)
self.logger.debug("Switch to primary: %s" % cmd)
(status, output) = commands.getstatusoutput(cmd)
if status != 0:
self.logger.logExit("Switch instance to be primary failed!Datadir %s.\nOutput: %s" % (datadir, output))
def doRestore(self):
"""
Restore the status of instances
"""
self.logger.log("Begin to restore instance status...")
try:
self.readConfigInfo()
self.getUserInfo()
# dump status to file
cmd = ClusterCommand.getQueryStatusCmd(self.user, self.dbNodeInfo.id, self.__curStatusFile)
(status, output) = commands.getstatusoutput(cmd)
if status != 0:
self.logger.logExit("Query local instance status failed!Error: %s" % output)
bakDbStatus = DbClusterStatus()
bakDbStatus.initFromFile(self.__bakStatusFile)
bakNodeStatus = bakDbStatus.getDbNodeStatusById(self.dbNodeInfo.id)
if bakNodeStatus is None:
self.logger.logExit("Get backup status of local node failed!")
curDbStatus = DbClusterStatus()
curDbStatus.initFromFile(self.__curStatusFile)
curNodeStatus = curDbStatus.getDbNodeStatusById(self.dbNodeInfo.id)
if curNodeStatus is None:
self.logger.logExit("Get current status of local node failed!")
if not curNodeStatus.isNodeHealthy():
self.logger.logExit("Current status of node is not healthy!")
# Compare the status and restore it
bakInstances = bakNodeStatus.datanodes + bakNodeStatus.gtms
for bakInst in bakInstances:
curInst = curNodeStatus.getInstanceByDir(bakInst.datadir)
if curInst is None:
self.logger.logExit("Get current status of instance failed!DataDir:%s" % bakInst.datadir)
if bakInst.status == curInst.status:
continue
if bakInst.status == DbClusterStatus.INSTANCE_STATUS_PRIMARY:
self.__switchToPrimary(bakInst.datadir)
elif bakInst.status == DbClusterStatus.INSTANCE_STATUS_STANDBY:
self.__switchToStandby(bakInst.datadir)
except Exception, e:
self.logger.logExit(str(e))
def doBackup(self):
"""
Backup the status of instances
"""
self.logger.log("Begin to backup instance status...")
try:
self.readConfigInfo()
self.getUserInfo()
# dump status to file
cmd = ClusterCommand.getQueryStatusCmd(self.user, self.dbNodeInfo.id, self.__bakStatusFile)
(status, output) = commands.getstatusoutput(cmd)
if status != 0:
self.logger.logExit("Query local instance status failed!Error: %s" % output)
except Exception, e:
self.logger.logExit(str(e))
if (user == ""):
GaussLog.exitWithError("Parameter input error, need '-U' parameter.")
if (logFile == ""):
logFile = DefaultValue.getOMLogPath(DefaultValue.DEFAULT_LOG_FILE, user, "")
writePid(user)
global g_logger
g_logger = GaussLog(logFile)
if (time_out <= 0):
time_out = 1800
endTime = datetime.now() + timedelta(seconds=time_out)
connList = ClusterCommand.readCooConnections(user)
if (len(connList) == 0):
raise Exception("There is no coordinator to connect!")
ip = connList[0][0]
port = connList[0][1]
sql = "select case (select pgxc_lock_for_backup()) when true then (select pg_sleep(%d)) end;" % time_out
cmd = "gsql -h %s -p %d postgres -X -c \"%s\"" % (ip, port, sql)
cmd = "su - %s -c '%s'" % (user, cmd)
sqlThread = CommandThread(cmd)
sqlThread.start()
time.sleep(5)
while True:
td = endTime - datetime.now()
leftSeconds = (td.microseconds + (td.seconds + td.days * 24 * 3600) * 10**6) / 10**6
if (leftSeconds <= 0):