python實現磁碟日誌清理的示例
阿新 • • 發佈:2020-11-06
一、描述:
以module的方式元件python程式碼,在磁碟檔案清理上覆用性更好
二、達到目標:
清空過期日誌檔案,清理掉超過自定大小日誌檔案
三、原碼
#!/usr/bin/env python # -*- coding: utf-8 -*- import commands import os import time import re import getopt import sys # commands.getstatusoutput 返回兩個元素的元組tuple(status,result),status為int型別,result為string型別 def execute_local_shell_cmd(cmd): status,result = commands.getstatusoutput(cmd) result = result.split("\n") return status,result def send_alert_mail(): pass ''' 獲取某一磁碟的空間使用率 ''' def get_disk_used(disk_name): status,result = execute_local_shell_cmd("df | grep %s | awk '{print $5}'" % disk_name) return status,result[0] #print(get_disk_used('/data0')) ''' 判斷檔案是否在指定時間內修改過 ''' def file_modify_in(file_path,time_interval='1d'): current_time = time.time() # os.path.getmtime 返回最後修改時間。返回從unix紀元開始的跳秒數 if current_time - os.path.getmtime(file_path) < translate_time_interval_to_second(time_interval): return True return False def translate_file_size_to_kb(file_size): # 將字串所有大寫字元轉為小寫 file_size = str(file_size.lower()) # 建立匹配數字1次或多次的數字且小數點出現一次或者不出現的;小數點後數字重複0次或多次模式物件 pattern = re.compile(r'\d+\.?\d*') match = pattern.match(file_size) file_size_number = None if match: # 使用Match獲得分組資訊 #print(match.group()) file_size_number = float(match.group()) else: raise IOError("Input {0} can't translate to byte." "Current support g(gb)/m(mb)/k(kb)/b(byte)".format(file_size)) # endswith() 方法用於判斷字串是否以指定字尾結尾,如果以指定字尾結尾返回True,否則返回False。 # 可選引數"start"與"end"為檢索字串的開始與結束位置。 if file_size.endswith("g") or file_size.endswith("gb"): return file_size_number * 1024 * 1024 * 1024 elif file_size.endswith("m") or file_size.endswith("mb"): return file_size_number * 1024 * 1024 elif file_size.endswith("k") or file_size.endswith("kb"): return file_size_number * 1024 elif file_size.endswith("b") or file_size.endswith("byte"): return file_size_number else: raise IOError("Input {0} can't translate to byte." "Current support g(gb)/m(mb)/k(kb)/b(byte)".format(file_size)) #print(translate_file_size_to_kb('10g')) def translate_time_interval_to_second(time_interval): date_interval = str(time_interval.lower()) pattern = re.compile(r'\d+') match = pattern.match(date_interval) date_interval_number = None if match: date_interval_number = int(match.group()) else: raise IOError("Input {0} can't translate to second." "Current support d(day)/h(hour)/m(min)/s(sec)".format(date_interval)) if date_interval.endswith('d') or date_interval.endswith('day'): return date_interval_number * 24 * 3600 elif date_interval.endswith('h') or date_interval.endswith('hour'): return date_interval_number * 3600 elif date_interval.endswith('m') or date_interval.endswith('min'): return date_interval_number * 60 elif date_interval.endswith('s') or date_interval.endswith('sec'): return date_interval_number else: raise IOError("Input {0} cant't translate to second." "Current support d(day)/h(hour)/m(min)/s(second)".format(date_interval)) #print(translate_time_interval_to_second('7d')) ''' 關斷檔案是否可能是當前log檔案 1) 修改改時間1天內 2) 以pattern結尾 ''' def probable_current_log_file(file_path,pattern='log',modify_in='1d'): if file_modify_in(file_path,time_interval=modify_in): return True return str(file_path).endswith(pattern) ''' 獲取超過天數設定log,注意不會返回可能是當前正在修改的檔案,檢視probable_current_log_file 確定如何做該判斷 ''' def get_clean_log_list_by_date(target_dir,before_days_remove='7d',pattern="log"): before_seconds_remove = translate_time_interval_to_second(before_days_remove) current_time = time.time() # os.listdir 返回指定資料夾包含檔案或資料夾的名字列表 for candidate_file in os.listdir(target_dir): candidate_file_fullpath = "%s/%s" %(target_dir,candidate_file) # 是否存在一個普通檔案 if os.path.isfile(candidate_file_fullpath): candidate_file_mtime = os.path.getmtime(candidate_file_fullpath) # find\(\)根據是否包含字串,如果包含有,返回開始的索引值,否則返回-1 if current_time - candidate_file_mtime > before_seconds_remove \ and candidate_file.find(pattern) != -1 \ and not probable_current_log_file(candidate_file_fullpath): # yield 就是return一個值,並且記住這個返回值的位置,下次迭代就從這個位置後開始 yield candidate_file_fullpath ''' 獲取超過大小的日誌檔案(注意預設不會返回修改時間小於1天的檔案) ''' def get_clean_log_list_by_size(target_dir,file_size_limit='10g',pattern="log"): file_size_limit_byte = translate_file_size_to_kb(file_size_limit) for candidate_file in os.listdir(target_dir): candidate_file_fullpath = "%s/%s" %(target_dir,candidate_file) if os.path.isfile(candidate_file_fullpath): # stat返回相關檔案的系統狀態資訊 file_stat = os.stat(candidate_file_fullpath) if candidate_file.find(pattern) != -1 and \ file_stat.st_size >= file_size_limit_byte: yield candidate_file_fullpath # 如果檔案在modify_in之內修改過,則不返回 # if not (modify_in and file_modify_in(candidate_file_fullpath,time_interval=modify_in)) and \ # not probable_current_log_file(candidate_file_fullpath): # yield candidate_file_fullpath ''' remove檔案列表 ''' def remove_file_list(file_list,roll_back=False): for file_item in file_list: if roll_back or probable_current_log_file(file_item,pattern=pattern,modify_in='1d'): print('roll back file %s' % file_item) execute_local_shell_cmd("cat /dev/null > {0}".format(file_item)) else: print('remove file %s' % file_item) # os.remove 刪除指定路徑檔案。如果指定的路徑是一個目錄,將丟擲OSError os.remove(file_item) ''' 清理掉超過日期的日誌檔案 ''' def remove_files_by_date(target_dir,pattern='log'): file_list = get_clean_log_list_by_date(target_dir,before_days_remove,pattern) remove_file_list(file_list) ''' 清理掉超過大小的日誌檔案 ''' def remove_files_by_size(target_dir,pattern='log'): file_list = get_clean_log_list_by_size(target_dir,file_size_limit,pattern) remove_file_list(file_list) ''' 清空當前的日誌檔案,使用cat /dev/null > {log_file}方式 ''' def clean_curren_log_file(target_dir,pattern='log'): for candidate_file in os.listdir(target_dir): candidate_file_fullpath = '%s/%s' % (target_dir,candidate_file) if candidate_file.endswith(pattern) and os.path.isfile(candidate_file_fullpath): file_stat = os.stat(candidate_file_fullpath) if file_stat.st_size >= translate_file_size_to_kb(file_size_limit): remove_file_list([candidate_file_fullpath],roll_back=True) def clean_data_release_disk(disk_name,target_dir,disk_used_limit='80%',pattern='log'): disk_used_limit = disk_used_limit.replace('%','') # 第一步執行按時間的日誌清理 print('Step one remove files {0} ago.'.format(before_days_remove)) remove_files_by_date(target_dir,before_days_remove=before_days_remove,pattern=pattern) # 如果磁碟空間還是沒有充分釋放,則執行按大小的日誌清理 current_disk_used = int(get_disk_used(disk_name)[1].replace('%','')) if current_disk_used > int(disk_used_limit): print("Disk {0}'s current used {1}% great than input used limit {2}%," "so we will remove files bigger than {3}". format(disk_name,current_disk_used,disk_used_limit,file_size_limit)) remove_files_by_size(target_dir,file_size_limit=file_size_limit,pattern=pattern) # 如果磁碟空間開沒有釋放,清空當前正在寫的log檔案,並alert current_disk_used = int(get_disk_used(disk_name)[1].replace('%'," "so we will roll back current log file". format(disk_name,file_size_limit)) clean_curren_log_file(target_dir,pattern=pattern) # 如果還是沒有,alert mail if int(get_disk_used(disk_name)[1].replace('%','')) > int(disk_used_limit): send_alert_mail() def usage(): print('clean.py -d <target_disk> -r <target_dirctory -u <diskUsedLimit(default 80%)> ' '-f <fileSizeLimit(default 10gb,gb/mb/kb)> -p <filePattern(default log)> ' '-t <beforeDaysRemove(default 7d,d)> ') if __name__ == "__main__": target_disk_input = '/data0' target_dir_input = '/data0/hadoop2/logs' disk_used_limit_input = '80%' file_size_limit_input = '10g' pattern_input = 'log' before_days_remove_input = '7d' try: # getopt 命令解析,有短選項和長選項 # getopt 返回兩人個引數:一個對應引數選項和value元組,另一個一般為空 opts,args = getopt.getopt(sys.argv[1:],'hd:r:u:f:p:t:',['help' 'disk=','directory=','diskUsedLimit=','fileSizeLimit=','filePattern=','beforeDaysRemove=']) # getopt模組函式異常錯誤,捕獲異常並列印錯誤 except getopt.GetoptError as err: print err usage() sys.exit(2) if len(opts) < 6: usage() sys.exit(2) for opt,arg in opts: if opt == '-h': usage() sys.exit() elif opt in ("-d","--disk"): target_disk_input = arg.replace('/','') elif opt in ("-r","--directory"): target_dir_input = arg elif opt in ("-u","--diskUsedLimit"): disk_used_limit_input = arg elif opt in ("-f","--fileSizeLimit"): file_size_limit_input = arg translate_file_size_to_kb(file_size_limit_input) elif opt in ("-p","filePattern"): pattern_input = arg elif opt in ("-t","--beforeDaysRemove"): before_days_remove_input = arg translate_time_interval_to_second(before_days_remove_input) print ("{0} Start clean job.target_disk:{1},target_directory:{2},disk_used_limit:{3}," "file_size_limit:{4},pattern:{5},before_days_remove:{6}".format(time.ctime(time.time()),target_disk_input,target_dir_input,disk_used_limit_input,file_size_limit_input,pattern_input,before_days_remove_input)) clean_data_release_disk(target_disk_input,disk_used_limit=disk_used_limit_input,file_size_limit=file_size_limit_input,pattern=pattern_input,before_days_remove=before_days_remove_input)
四、統一呼叫目錄定時刪除
#!/usr/bin/env python # -*- coding: utf-8 -*- import os # 遍歷目錄 def Lisdir(targetdir): list_dirs = os.walk(targetdir) for root,list_dirs,files in list_dirs: for d in list_dirs: yield os.path.join(root,d) def log_dir(targetdir): list_dirs = os.listdir(targetdir) for ph in list_dirs: if os.path.isdir(os.path.join(targetdir,ph)): yield Lisdir(os.path.join(targetdir,ph)) for path in log_dir('/data0/backup_log-bin'): for ppp in path: # 以log-bin結尾 為假 if ppp.endswith('log-bin') is False: os.system("db_script/clean_robo.py -d /data0 -r {0} -u 75% -f 501M -p bin -t 5d".format(ppp))
以上就是python實現磁碟日誌清理的示例的詳細內容,更多關於python 磁碟日誌清理的資料請關注我們其它相關文章!