def log_error():
req = request.values
page = int(req.get("p", 1))
date_from = req.get("date_from", DateHelper.getCurrentTime(fmt="%Y-%m-%d"))
date_to = req.get("date_to", DateHelper.getCurrentTime(fmt="%Y-%m-%d"))
query = AppErrLog.query.filter(
AppErrLog.created_time.between(date_from, date_to + " 23:59:59"))
page_params = {
"total": query.count(),
"page_size": CommonConstant.PAGE_SIZE,
"page": page,
"display": CommonConstant.PAGE_DISPLAY,
}
pages = UtilHelper.iPagination(page_params)
offset = (page - 1) * CommonConstant.PAGE_SIZE
limit = CommonConstant.PAGE_SIZE * page
list = query.order_by(AppErrLog.id.desc())[offset:limit]
sc = {'date_from': date_from, 'date_to': date_to}
return UtilHelper.renderView("home/log/error.html", {
"list": list,
"pages": pages,
"sc": sc
})
def tools_alert():
req = request.values
page = int(req.get("p", 1))
date_from = req.get("date_from", DateHelper.getCurrentTime(fmt="%Y-%m-%d"))
date_to = req.get("date_to", DateHelper.getCurrentTime(fmt="%Y-%m-%d"))
status = int(req.get("status", CommonConstant.default_status_neg_99))
query = JobAlertList.query.filter(
JobAlertList.created_time.between(date_from, date_to + " 23:59:59"))
if status > CommonConstant.default_status_neg_99:
query = query.filter_by(status=status)
page_params = {
"total": query.count(),
"page_size": CommonConstant.PAGE_SIZE,
"page": page,
"display": CommonConstant.PAGE_DISPLAY,
}
pages = UtilHelper.iPagination(page_params)
offset = (page - 1) * CommonConstant.PAGE_SIZE
limit = CommonConstant.PAGE_SIZE * page
list = query.order_by(JobAlertList.id.desc())[offset:limit]
data = []
if list:
job_ids = ModelHelper.getFieldList(list, "job_id")
job_map = ModelHelper.getDictFilterField(JobList,
select_field=JobList.id,
id_list=job_ids.sort())
for item in list:
tmp_data = ModelHelper.model2Dict(item)
tmp_job_info = ModelHelper.model2Dict(
job_map.get(tmp_data['job_id']))
tmp_data['status_desc'] = CommonConstant.common_status_map4[
tmp_data['status']]
tmp_data['job_name'] = tmp_job_info['name']
data.append(tmp_data)
sc = {'date_from': date_from, 'date_to': date_to, 'status': status}
return UtilHelper.renderView(
"home/job/tools/alert.html", {
"list": data,
"pages": pages,
"sc": sc,
"status_map": CommonConstant.common_status_map4,
"current": "alert"
})
def home_index():
date = DateHelper.getCurrentTime(fmt="%Y-%m-%d")
job_count = JobList.query.filter_by(is_del=CommonConstant.default_status_false).count()
server_count = JobServer.query.filter_by(status=CommonConstant.default_status_true).count()
alert_count = JobAlertList.query.filter(JobAlertList.created_time.between(date, date + " 23:59:59")).count()
cate_map = ModelHelper.getDictFilterField( JobCategory )
cat_job_map = {}
cat_job_list = JobList.query.with_entities(JobList.cate_id ,func.count( JobList.id) )\
.filter_by(is_del=CommonConstant.default_status_false)\
.group_by(JobList.cate_id).all()
if cat_job_list:
for _item in cat_job_list:
cat_job_map[ _item[0] ] = _item[1]
type_job_map = {}
job_type_map = CommonConstant.job_type_map
type_job_list = JobList.query.with_entities(JobList.job_type, func.count(JobList.id)) \
.filter_by(is_del=CommonConstant.default_status_false) \
.group_by(JobList.job_type).all()
if type_job_list:
for _item in type_job_list:
type_job_map[ _item[0] ] = _item[1]
return UtilHelper.renderView("home/index/index.html", {
"job_count": job_count,
"server_count": server_count,
"alert_count": alert_count,
'cate_map' : cate_map,
'cat_job_map' : cat_job_map,
'job_type_map' : job_type_map,
'type_job_map' : type_job_map,
})
def getReleaseVersion():
ver = "%s" % (DateHelper.getCurrentTime("%Y%m%d%H%M%S%f"))
release_path = app.config.get('RELEASE_PATH')
if release_path and os.path.exists(release_path):
with open(release_path, 'r') as f:
ver = f.readline()
return ver
def job_info():
req = request.values
id = int(req['id']) if ('id' in req and req['id']) else 0
info = JobList.query.filter_by(id=id).first()
if not info:
return redirect(GlobalUrlService.buildHomeUrl("/job/index/index"))
info = ModelHelper.model2Dict(info)
server_info = JobServer.query.filter_by(id=info['server_id']).first()
cate_info = JobCategory.query.filter_by(id=info['cate_id']).first()
server_env_map = CommonConstant.server_env_map
run_status_map = CommonConstant.run_status_map
info['next_run_time'] = DateHelper.getDateOnTimestamps(
info['next_run_time'], '%Y-%m-%d %H:%M')
info['env_name'] = server_env_map.get(info['env_id'])
info['run_status_desc'] = run_status_map.get(info['run_status'])
info['job_status_desc'] = job_status_map.get(info['status'])
info['server_name'] = server_info.name
info['cate_name'] = cate_info.name if cate_info else ''
info['run_interval_desc'] = DateHelper.formatBeautyTime(
info['run_interval'] * 60)
user_map = ModelHelper.getDictFilterField(
User,
select_field=User.id,
id_list=[info['owner_uid'], info['relate_uid']])
##获取最近5天运行记录
log_list = JobRunLog.query.filter_by(job_id=id).order_by(
JobRunLog.id.desc())[0:5]
log_data = []
if log_list:
for item in log_list:
tmp_data = ModelHelper.model2Dict(item)
tmp_data['status_desc'] = CommonConstant.job_log_status_map[
tmp_data['status']]
tmp_data['duration'] = ""
if DateHelper.getCurrentTime(date=tmp_data['end_time']
) == CommonConstant.DEFAULT_DATETIME:
tmp_data['end_time'] = "未知"
tmp_data['duration'] = time.time() - time.mktime(
tmp_data['start_time'].timetuple())
else:
tmp_data['duration'] = tmp_data['end_time'].timestamp(
) - tmp_data['start_time'].timestamp()
tmp_data['duration'] = DateHelper.formatBeautyTime(
tmp_data['duration'])
log_data.append(tmp_data)
return UtilHelper.renderView(
"home/job/index/info.html", {
"info": info,
"log_list": log_data,
"user_map": user_map,
"job_level_map": CommonConstant.job_level_map,
})
def updateRunLog(log_id=0, max_mem=0, status=0):
params = {
"end_time": DateHelper.getCurrentTime(),
"max_mem": max_mem,
"status": status
}
JobRunLog.query.filter_by(id=log_id).update(dict(params))
db.session.commit()
return True
def index():
return "APP_NAME：{0} VERSION：{1} APP：API {2}"\
.format( app.config.get("APP_NAME"),app.config.get("APP_VERSION"), DateHelper.getCurrentTime() )
def tools_log():
status_map = CommonConstant.job_log_status_map
req = request.values
page = int(req.get("p", 1))
date_from = req.get("date_from", DateHelper.getCurrentTime(fmt="%Y-%m-%d"))
date_to = req.get("date_to", DateHelper.getCurrentTime(fmt="%Y-%m-%d"))
status = int(req.get("status", CommonConstant.default_status_neg_99))
job_id = int(req.get("job_id", CommonConstant.default_status_false))
query = JobRunLog.query.filter(
JobRunLog.created_time.between(date_from, date_to + " 23:59:59"))
if job_id:
query = query.filter_by(job_id=job_id)
if status > CommonConstant.default_status_neg_99:
query = query.filter_by(status=status)
page_params = {
"total": query.count(),
"page_size": CommonConstant.PAGE_SIZE,
"page": page,
"display": CommonConstant.PAGE_DISPLAY,
}
pages = UtilHelper.iPagination(page_params)
offset = (page - 1) * CommonConstant.PAGE_SIZE
limit = CommonConstant.PAGE_SIZE * page
list = query.order_by(JobRunLog.id.desc())[offset:limit]
data = []
if list:
job_ids = ModelHelper.getFieldList(list, "job_id")
job_map = ModelHelper.getDictFilterField(JobList,
select_field=JobList.id,
id_list=job_ids.sort())
for item in list:
tmp_data = ModelHelper.model2Dict(item)
tmp_job_info = ModelHelper.model2Dict(
job_map.get(tmp_data['job_id']))
tmp_data['status_desc'] = status_map[tmp_data['status']]
tmp_data['job_name'] = tmp_job_info['name']
tmp_data['duration'] = ""
if DateHelper.getCurrentTime(date=tmp_data['end_time']
) == CommonConstant.DEFAULT_DATETIME:
tmp_data['end_time'] = "未知"
tmp_data['duration'] = time.time() - time.mktime(
tmp_data['start_time'].timetuple())
else:
tmp_data['duration'] = (tmp_data['end_time'].timestamp() -
tmp_data['start_time'].timestamp())
tmp_data['duration'] = DateHelper.formatBeautyTime(
tmp_data['duration'])
data.append(tmp_data)
sc = {
'date_from': date_from,
'date_to': date_to,
'status': status,
'job_id': job_id
}
return UtilHelper.renderView(
"home/job/tools/log.html", {
"list": data,
"pages": pages,
"sc": sc,
"status_map": status_map,
"current": "log"
})
def run(self, params):
pid_path = self.getPidPath('dispatch.pid')
if self.checkPidExist(pid_path):
app.logger.info("[core] dispatch is running")
return False
pid = str(os.getpid())
if not self.setPidFile(pid_path, pid):
err_msg = self.getErrMsg()
app.logger.info("Cann't get a lock file，err msg : " + err_msg)
return False
params = self.getEnvFile()
server_id = params['id']
host = params['name']
list = JobList.query.filter_by( server_id = server_id,status = CommonConstant.default_status_true,
run_status = CommonConstant.default_status_true,is_del = CommonConstant.default_status_false ).all()
if not list:
app.logger.info("没有数据需要调度~~")
return True
for t in list:
##调度时间是否到了，应该加入到SQL查询中查询
if t.next_run_time >= time.time():
app.logger.info( "job_id:%s，运行时间未到"%( t.id ) )
continue
# 启子进程中运行
app.logger.info( "get a task: job_id:%s，运行时间：%s" % (t.id,DateHelper.getCurrentTime()) )
'''
子进程复制一份父进程内存给自己，两个进程之间的执行是相互独立的，其执行顺序可以是不确定的、随机的、不可预测的
多进程（os.fork()）下，子进程会继承父进程的连接，所以会有问题.先销毁已有的engine，确保父进程没有数据库连接
相关错误：Mysql server has gone away
'''
self.closePoolDB()
pid = os.fork()
if pid == 0: #子进程,这里是一个独立进程（在复制出来的那一刻 所有变量都会共享到子进程）,所以写代码 就要感觉在一个独立方法中
self.closePoolDB()
job_id = t.id
job_pid_file = self.getPidPath( 'job_%s.pid' % job_id )
if self.checkPidExist(job_pid_file):
app.logger.info("job_id:%s is running on %s" % (job_id, host))
return 0
## 建立具体job的pid 文件，防止重复运行，这个进程id其实记录的是不对的，应该使用下面的
tmp_pid = str(os.getpid())
if not self.setPidFile(job_pid_file, tmp_pid):
app.logger.info("job_id:%s 不能建立pid，path：%s，msg：%s" % (job_id, job_pid_file, self.getErrMsg()))
return True
app.logger.info("job_id:%s 建立pid，子进程pid：%s" % (job_id, tmp_pid))
## 更新job为运行中
try:
tmp_affect_rows = JobList.query.filter_by( id = job_id,run_status = CommonConstant.default_status_true )\
.update( dict( run_status = CommonConstant.default_status_pos_2 ) )
db.session.commit()
if tmp_affect_rows < 1:
app.logger.info("job_id:%s不能得到lock，任务已经运行中" % job_id)
return False
except:
app.logger.info( "job_id:%s不能得到锁状态，err：%s" % (job_id,str( sys.exc_info() ) ) )
##写入一条调度日志
tmp_log_id = 0
try:
tmp_log_params = {
"job_id":job_id,
"server_id":server_id,
"server_name":host,
"status": CommonConstant.default_status_neg_1,
"start_time":DateHelper.getCurrentTime()
}
tmp_log_id = JobService.addRunLog( tmp_log_params )
except :
pass
self.closePoolDB()
tmp_job_run_start_time = time.time() # job开始运行的时间
# t.command无法获取job内部输出的内容，我们需要按行读取或者按buffer读取的
# status = os.system(t.command)>>8#status, output = commands.getstatusoutput(t.command)
# 可以加 bufsize = -1 表示使用系统默认缓存
# 命令前面加一下前缀，方便搜索关键词查找
# 创建子进程后子进程不结束 https://bbs.csdn.net/topics/390596479?_=1515055076
tmp_command = t.command
tmp_command = "tmp_job_%s='%s' && %s"%( job_id,DateHelper.getCurrentTime(),tmp_command)
#如果想要达到2>&1 可以设置为stdout=subprocess.PIPE,stderr=subprocess.STDOUT
sp = subprocess.Popen(tmp_command, bufsize = -1, shell = True, close_fds=True, stdout=subprocess.PIPE,stderr=subprocess.STDOUT)
tmp_run_job_pid = sp.pid
#上面存的关于job的进程id 是不对的，这里在覆盖一次
self.coverPidFile(job_pid_file,tmp_run_job_pid)
'''
如果是包裹了一层runshell的需要找到进程的子进程pid，然后在查看内存
tmp_pid 是目前子进程的进程号
tmp_run_job_pid 是目前 subprocess启动的进程好
真真运行的进程号，需要通过关键词查询，所以在命令前面加了特别东西
'''
app.logger.info( "job_id:%s 启动Job脚本，pid：%s" % (job_id, tmp_run_job_pid))
##统计内存占用量
tmp_max_job_used_mem = tmp_job_used_mem = UtilHelper.getUsedMemory( tmp_run_job_pid )
app.logger.info("job_id:%s 日志start-------" %job_id )
'''
sp.poll() 返回值
0 正常结束
1 sleep
2 子进程不存在
-15 kill
None 在运行
判断进程状态
'''
while sp.poll() is None:
##统计内存占用量
tmp_job_used_mem = UtilHelper.getUsedMemory( tmp_run_job_pid )
if tmp_job_used_mem > tmp_max_job_used_mem:
tmp_max_job_used_mem = tmp_job_used_mem
tmp_line_output = sp.stdout.readline()
tmp_line_output = tmp_line_output.strip()
#返回的是bytes
tmp_line_output = str(tmp_line_output, encoding="utf8")
tmp_lines = tmp_line_output.split("\n")
for tmp_line in tmp_lines:
app.logger.info("job_id:%s %s" % (job_id, tmp_line) )
app.logger.info("job_id:%s 日志end-------" % job_id)
##统计内存占用量
tmp_job_used_mem = UtilHelper.getUsedMemory( tmp_run_job_pid )
if tmp_job_used_mem > tmp_max_job_used_mem:
tmp_max_job_used_mem = tmp_job_used_mem
app.logger.info("job_id:%s PID:%s, 使用内存（end） %s" % (job_id, tmp_run_job_pid, tmp_job_used_mem))
app.logger.info("job_id:%s PID:%s, 最大使用内存 %s" % (job_id, tmp_run_job_pid, tmp_max_job_used_mem))
app.logger.info("job_id:%s 更新消耗内存完毕" % (job_id))
# 将标准输出关闭了
sp.stdout.close()
tmp_status = sp.wait()
app.logger.info("job_id:%s status_code:%s，%s" % (t.id, str(tmp_status),tmp_command ))
#和下面分开就是怕报警影响正常处理
try:
#杀死常驻job也会发生 MySQL server has gone away，只要运行的时间太长就会出问题了
self.closePoolDB()
#相关报警判断
self.alertStatusJudge(t, tmp_status)
self.alertRunTimeJudge(t, tmp_job_run_start_time)
except:
app.logger.info( self.getErrMsg() )
# 更新状态和下一次运行时间
try:
self.closePoolDB()
##提前将文件释放下，因为当服务器状态非常繁忙的时候，进程比较缓慢，会导致状态已经更新但是pid文件没有删除
self.atexit_removepid(job_pid_file)
if int( t.job_type ) == CommonConstant.default_status_pos_3 :#一次性job
JobList.query.filter_by(id=job_id).update( dict( run_status = CommonConstant.default_status_false,status = CommonConstant.default_status_false) )
db.session.commit()
else:
if int( t.job_type ) == CommonConstant.default_status_pos_2: # 常驻Job，他停止之后下一分钟直接运行
tmp_next_time = datetime.datetime.now() + datetime.timedelta(minutes=1)
tmp_next_time = tmp_next_time.replace(second=0)
tmp_next_time = int( time.mktime(tmp_next_time.timetuple() ) )
else:
tmp_next_time = t.next_run_time + int( math.ceil((time.time() - t.next_run_time) / (t.run_interval * 60)) * t.run_interval * 60)
JobList.query.filter_by(id=job_id).update( dict( run_status = CommonConstant.default_status_true ,next_run_time = tmp_next_time ) )
db.session.commit()
##更新对应日志的log,担心出错影响其他
JobService.updateRunLog(tmp_log_id, tmp_max_job_used_mem, (tmp_status == 0))
except:
app.logger.info( self.getErrMsg() )
# 完成
self.closePoolDB()
app.logger.info('job_id:%s 运行完成时间为：%s，子进程结束~~' % (job_id, DateHelper.getCurrentTime() ))
return 0
elif pid > 0: # 父进程
'''
status是一个传出参数。
waitpid的pid参数选择：
< -1 回收指定进程组内的任意子进程
= -1 回收任意子进程,等待所有的子进程终止
= 0 回收和当前调用waitpid一个组的所有子进程
> 0 回收指定ID的子进程
'''
app.logger.info("父进程 job_id:%s pid：%s" % (t.id, pid))
#os.waitpid( pid , os.WNOHANG)
os.waitpid(-1, os.WNOHANG)
app.logger.info("job_id:%s 父进程结束~~" % t.id)
else:
app.logger.info("job_id:%s,不能建立调度器" % (t.id))
app.logger.info("it's over~~")
return True