def pingLan():
if lanIP == dbOneIp and lanIP != dbTwoIp:
portOK = chkPort(dbTwoIp, int(3307))
if not portOK:
doPing(dbTwoIp, int(3307))
else:
if sDb.read(telIP, "lanAlarm") == "OFF": ## 状态OK且开关为"OFF"时,报恢复
sub = "DBTX LanIP is OK: %s" % ps1
msg = "DateTime: %s\n 备库: %s, 网络与端口恢复正常." % (getTimeNow(), dbTwoIp)
saveLog.info(msg)
sendMail(sub, msg)
sDb.update(telIP, "lanAlarm", "ON")
elif lanIP != dbOneIp and lanIP == dbTwoIp:
portOK = chkPort(dbOneIp, int(3306))
if not portOK:
doPing(dbOneIp, int(3306))
else:
if sDb.read(telIP, "lanAlarm") == "OFF":
sub = "DBTX LanIP is OK: %s" % ps1
msg = "DateTime: %s\n 备库: %s, 网络与端口恢复正常." % (getTimeNow(), dbTwoIp)
saveLog.info(msg)
sendMail(sub, msg)
sDb.update(telIP, "lanAlarm", "ON")
else:
pass
def getDisk(self):
mountPoints, diskTotal = [], 0 #diskTotal累加
parts = psutil.disk_partitions() #分区信息
for p in parts:
mp = str(p).split(",")[1].split("=")[1]
mountPoints.append(mp)
for x in mountPoints:
pTotal = psutil.disk_usage(x.strip("'")).total #单个分区容量
diskTotal += pTotal
diskTotal = changeUnit(int(diskTotal))
rootSize, rootRate = psutil.disk_usage('/').free, psutil.disk_usage('/').percent #获取/分区的剩余空间和使用率
writeToFile("totalDisk", diskTotal)
writeToFile("rootSize", changeUnit(int(rootSize)))
writeToFile("rootRate", rootRate)
if float(rootRate) > float(85):
"""ON开关，防止多次重复报警；报警一次之后设为关闭"""
if self.sDb.read(self.telIP, "rootAlarm") == "ON": ## 状态为"NO"且开关为"ON"时,报故障
sub = "DBTX ROOT partition Alarm:%s" % self.ps1
msg = "DateTime: %s\n根分区已使用%s%%, 超过85%%, 请警惕!!!" % (getTimeNow(), rootRate)
saveLog.warning(msg)
sendMail(sub, msg)
self.sDb.update(self.telIP, "rootAlarm", "OFF") ## 报警后,将开关置为"OFF"
else:
if self.sDb.read(self.telIP, "rootAlarm") == "OFF": ## 状态OK且开关为"OFF"时,报恢复
sub = "DBTX ROOT partition is OK:%s" % self.ps1
msg = "DateTime: %s\n根分区可用空间充足%s." % (getTimeNow(), changeUnit(int(rootSize)))
saveLog.info(msg)
sendMail(sub, msg)
self.sDb.update(self.telIP, "rootAlarm", "ON") ## 使用率不足85%时,将开关置为"ON"
def doPing(tIP, dbPort):
if dbPort == 3307:
mString = "备库-%s-3307端口不通" % tIP
else:
mString = "主库-%s-3306端口不通" % tIP
p = subprocess.Popen("ping -c7 %s" % tIP,shell=True,stdout=subprocess.PIPE,stderr=subprocess.PIPE)
pResult = p.stdout.read()
lossRate = pResult.split("---")[-1].split(",")[2].split()[0] ## 取丢包率
if float(lossRate[:-1]) > 20:
sub = "DBTX LanIP Alarm: %s" % ps1
msg = "DateTime: %s\n%s, ICMP-Ping丢包率:%s%% !!!" % (getTimeNow(), mString, lossRate)
else:
sub = "DBTX LanIP-Port Alarm: %s" % ps1
msg = "DateTime: %s\n%s, ICMP-Ping正常,可能是MySQL挂了或者防火墙对应端口没有开启 !!!" % (getTimeNow(), mString)
saveLog.error(msg)
if sDb.read(telIP, "lanAlarm") == "ON": ## 状态为"NO"且开关为"ON"时,报故障
sendMail(sub, msg)
sDb.update(telIP, "lanAlarm", "OFF") ## 报警后,将开关置为"OFF"
def getNginxStatus(self):
try:
if os.path.exists("/usr/local/nginx/logs/nginx.pid"):
fObj = open("/usr/local/nginx/logs/nginx.pid")
ngxMainPid = fObj.read().strip()
fObj.close()
else:
ngxMainPid = 0
pids = getProcessID("nginx")
if ngxMainPid:
if not pids or len(pids) < 8:
wp = ""
res = os.popen("ps -C nginx -o pid,cmd").readlines()
for l in res:
wp += "\n" + l.strip()
if self.sDb.read(telIP, "ngxAlarm") == "ON": ## 状态为"NO"且开关为"ON"时,报故障
sub = "DBTX Nginx Alarm: %s" % ps1
msg = "Datetime: %s\nNginx主进程PID:%s\nWorker_processes:%s\n少于10个!!" % (getTimeNow(), ngxMainPid, wp)
saveLog.warning(msg)
sendMail(sub, msg)
self.sDb.update(telIP, "ngxAlarm", "OFF")
else:
if self.sDb.read(telIP, "ngxAlarm") == "OFF": ## 状态OK且开关为"OFF"时,报恢复
sub = "DBTX Nginx is OK: %s" % ps1
msg = "Datetime: %s\nNginx主进程恢复开启状态." % (getTimeNow())
saveLog.info(msg)
sendMail(sub, msg)
self.sDb.update(telIP, "ngxAlarm", "ON")
else:
if self.sDb.read(telIP, "ngxAlarm") == "ON": ## 状态为"NO"且开关为"ON"时,报故障
sub = "DBTX Nginx Alarm: %s" % ps1
msg = "Datetime: %s\nNginx主进程挂了!!" % (getTimeNow())
saveLog.warning(msg)
sendMail(sub, msg)
self.sDb.update(telIP, "ngxAlarm", "OFF")
except Exception,err:
saveLog.warning("getNginxStatus, %s" % str(err))
def appCpuAnalyze(srvPidStatus, srvFlag, sDb):
from dbtx.Merge.etc import ps1
corMuser = "******"
tNowCpu = topNow()
if not sDb.exists(srvFlag, "ct"): ## Redis库中不存在该服区标识,初始化游戏程序占用CPU达到95%的次数
sDb.intAppCpuCt(srvFlag)
for app in srvPidStatus:
appName = app.split("_")[0]
if srvPidStatus[app][app+"_status"] == 0 or "Snapshot" in app: ## 如果app状态异常,不检测CPU,MEM信息,忽略掉Snapshot
continue
appPid = srvPidStatus[app][app+"_pid"].split()
for pid in appPid: ## CellApp,GateApp,CopyApp会有多个进程ID
nCpu, appCount = tNowCpu[pid], int(sDb.read(appName, srvFlag))
if float(nCpu) > 95.5: ## CPU使用率都大于95%时,count+1
appCount += 1
sDb.update(appName, srvFlag, appCount) ## 将新的count值更新到Redis库
else:
sDb.update(appName, srvFlag, 0) ## 如果没有超过95%,则置为0
if int(appCount) >= 8:
if sDb.read(srvFlag, "isFixing") == "OFF": ## 如果维护标识为"OFF"
if sDb.read(srvFlag, "appCpuAlarm") == "ON": ## 状态为"NO"且开关为"ON"时,报故障
sub = "DBTX App-CPU Alarm: %s" % ps1
msg = "DateTime: %s\n%s CPU使用率连续五次达到95%%, 进程ID为: %s !!!" % (getTimeNow(), app, pid)
saveLog.warning(msg)
sendMail(sub, msg, corMuser)
## sDb.updateDb(mulDb, srvFlag, "appCpuAlarm", "OFF") ## 报警后,将开关置为"OFF"
sDb.update(appName, srvFlag, 0) ## 报警后,重置count值
if "DbCacheApp_"+srvFlag == app:
proc = psutil.Process(int(pid))
shr = proc.get_ext_memory_info()[2]
if float(shr) > 1932735283.2: ## DbCacheApp共享内存大于1.8G时,报警
if sDb.read(srvFlag, "isFixing") == "OFF": ## 如果维护标识为"OFF"
if sDb.read(srvFlag, "appShrAlarm") == "ON": ## 状态为"NO"且开关为"ON"时,报故障
sub = "DBTX DbCacheApp-SHR Alarm: %s" % ps1
msg = "DateTime: %s\n%s DbCacheApp 共享内存达到红色警戒,当前占用共享内存: %s !!!" % (getTimeNow(), srvFlag, changeUnit(shr))
saveLog.warning(msg)
sendMail(sub, msg, corMuser)
sDb.update(srvFlag, "appShrAlarm", "OFF") ## 报警后,将开关置为"OFF"
else:
if sDb.read(srvFlag, "isFixing") == "OFF": ## 如果维护标识为"OFF"
if sDb.read(srvFlag, "appShrAlarm") == "OFF": ## 状态OK且开关为"OFF"时,报恢复
sub = "DBTX DbCacheApp-SHR Alarm: %s" % ps1
msg = "DateTime: %s\n%s DbCacheApp 占用共享内存恢复到正常水平: %s ." % (getTimeNow(), srvFlag, changeUnit(shr))
saveLog.warning(msg)
sendMail(sub, msg, corMuser)
sDb.update(srvFlag, "appShrAlarm", "ON")
def gateAnalyze(srvFlag, nowNu, sDb):
from dbtx.Merge.etc import ps1
pubFile = "/opt/dbtx/" + str(srvFlag) + "/config/public_config.xml"
pubObj = open(pubFile)
fixGateNum = ''.join([ l.strip().split(">")[1].split("<")[0] for l in pubObj.readlines() if "<GateNum>" in l])
pubObj.close()
corMuser = "******"
if int(nowNu) < int(fixGateNum):
if sDb.read(srvFlag, "isFixing") == "OFF": ## 如果维护标识为"OFF"
if sDb.read(srvFlag, "appNumAlarm") == "ON": ## 状态为"NO"且开关为"ON"时,报故障
sub = "DBTX GateApps Alarm: %s" % ps1
msg = "DateTime: %s\n%s 预设开启 %s 个GateApp,当前只有 %s 个,挂了 %s 个 !!!" % (getTimeNow(), srvFlag, fixGateNum, nowNu, (int(fixGateNum)-int(nowNu)))
saveLog.warning(msg)
sendMail(sub, msg, corMuser)
sDb.update(srvFlag, "appNumAlarm", "OFF") ## 报警后,将开关置为"OFF"
else:
if sDb.read(srvFlag, "isFixing") == "OFF": ## 如果维护标识为"OFF"
if sDb.read(srvFlag, "appNumAlarm") == "OFF": ## 状态OK且开关为"OFF"时,报恢复
sub = "DBTX GateApps is OK: %s" % ps1
msg = "DateTime: %s\n%s GateApp Numbers正常." % (getTimeNow(), srvFlag)
saveLog.warning(msg)
sendMail(sub, msg, corMuser)
sDb.update(srvFlag, "appNumAlarm", "ON")
redisData["connects_rate"] = (int(redisData["total_connects"]) - int(last_connects)) / float(60)
redisData["command_rate"] = (int(redisData["total_commands"]) - int(last_commands)) / float(60)
fZbx = str(redisData).strip("{}").replace(",","\n").replace("'","") ## 转换成zabbix需要的格式
fObj = open(rZbx,'w')
fObj.write(fZbx)
fObj.close()
except Exception,err:
saveLog.error("getRedisBaseInfo(): %s" % str(err))
## 分析数据,必要时报警: 1, 使用内存大于1G; 2, 命中率低于20%
if int(redisData["used_memory"]) > 1073741824:
sDb = opRedis() ## 实例化 Redis库操作类
if sDb.read(telIP, "rRateAlarm") == "ON":
sub = "DBTX Redis Memory Alarm: %s" % ps1
msg = "DateTime: %s\nRedis使用内存为: %s, 超过1G,请警惕 !!!" % (getTimeNow(), changeUnit(int(redisData["used_memory"])))
saveLog.warning(msg)
sendMail(sub, msg)
sDb.update(telIP, "rRateAlarm", "OFF") ## 报警后,将开关置为"OFF"
else:
sDb = opRedis() ## 实例化 Redis库操作类
if sDb.read(telIP, "rRateAlarm") == "OFF":
sub = "DBTX Redis Memory is OK: %s" % ps1
msg = "DateTime: %s\nRedis使用内存情况正常: %s." % (getTimeNow(), changeUnit(int(redisData["used_memory"])))
saveLog.info(msg)
sendMail(sub, msg)
sDb.update(telIP, "rRateAlarm", "ON")
## if totalHits != 0 and redisData["hit_rate"] < 20:
## sDb = opShelveDb() ## 实例化 shelve库操作类
## if sDb.readDb(pubDb, telIP, "rRateAlarm") == "ON":
def getServerTime(self):
nowTime = getTimeNow()
self.data_for_post += "server_time=%s&" % nowTime
nowTime = getTimeNow()
self.data_for_post += "server_time=%s&" % nowTime
def getDbStatus(self):
## 检查MySQL状态, 1 正常, 0 异常
try:
db1Status = chkPort(dbOneIp, db1Port)
db2Status = chkPort(dbTwoIp, db2Port)
except Exception, err:
saveLog.error("getDbStatus() Error Occur : %s" % err)
if db1Status:
db1_status = 1
if self.sDb.read(telIP, "db1Alarm") == "OFF": ## 状态OK且开关为"OFF"时,报恢复
sub = "DBTX DB1 mysql is OK: %s" % ps1
msg = "DateTime: %s\n主库-%s-3306端口 is up" % (getTimeNow(), dbOneIp)
saveLog.info(msg)
sendMail(sub, msg)
self.sDb.update(telIP, "db1Alarm", "ON")
## statusAnalyze("3306") ## 获取其它性能数据,zabbix成图或报警
else:
db1_status = 0
try:
if self.sDb.read(telIP, "db1Alarm") == "ON": ## 状态为"NO"且开关为"ON"时,报故障
sub = "DBTX DB1 mysql Alarm: %s" % ps1
msg = "DateTime: %s\n主库-%s-3306端口 DOWN 了,请查看日志 !!!" % (getTimeNow(), dbOneIp)
saveLog.warning(msg)
sendMail(sub, msg)
self.sDb.update(telIP, "db1Alarm", "OFF") ## 报警后,将开关置为"OFF"
except Exception, err:
saveLog.error("db1Alarm() Error Occur : %s" % str(err))
if fixSwStat == "OFF": ## 如果维护标识为"OFF"
if appSwStat == "ON": ## 状态为"NO"且开关为"ON"时,报故障
sub = "DBTX AppService Alarm: %s" % ps1
msg = "DateTime: %s\n%s 挂啦 !!!" % (getTimeNow(), srvStatus['Error'])
saveLog.warning(msg)
sendMail(sub, msg, mUser)
self.sDb.update(srvFlag, "appAlarm", "OFF")
except Exception, err:
saveLog.warning("app_status is 2,get switch status error occured.")
elif srvStatus['Error'] and not srvStatus['Success']: ## 全部处于关闭状态
app_status = 0
try:
if fixSwStat == "OFF": ## 如果维护标识为"OFF"
if appSwStat == "ON": ## 状态为"NO"且开关为"ON"时,报故障
sub = "DBTX AppService is Down: %s" % ps1
msg = "DateTime: %s\n%s AppService 关闭 !!!" % (getTimeNow(), srvFlag)
saveLog.warning(msg)
sendMail(sub, msg, mUser)
self.sDb.update(srvFlag, "appAlarm", "OFF")
except Exception, err:
saveLog.warning("app_status is 0,get switch status error occured.")
elif not srvStatus['Error'] and srvStatus['Success']: ## 全部处于开启状态
app_status = 1
appCpuAnalyze(srvPidStatus,srvFlag,self.sDb)
self.getSrvVer(srvPidStatus)
gateAnalyze(srvFlag,GateAppCount,self.sDb)
try:
if fixSwStat == "OFF": ## 如果维护标识为"OFF"
if appSwStat == "OFF": ## 状态OK且开关为"OFF"时,报恢复
sub = "DBTX AppService is OK: %s" % ps1
msg = "DateTime: %s\n%s AppService 恢复开启 !!!" % (getTimeNow(), srvFlag)
