Structured Streaming checkpoint

checkpoint主要是面向kafkaStreamSource锦援、rateStreamSource

/commits
/metadata
/offsets
/sources
/state

1、commitlog內(nèi)容如下:
/commits/262
/commits/263

hadoop fs -cat /commits/164
v1
{}
注:version with prefix v

2曼库、offsetSeq

/offsets/263
/offsets/264

hadoop fs -cat /offsets/265
v1
{"batchWatermarkMs":0,"batchTimestampMs":1533975004482,"conf":{"spark.sql.shuffle.partitions":"200","spark.sql.streaming.stateStore.providerClass":"org.apache.spark.sql.execution.streaming.state.HDFSBackedStateStoreProvider"}}
{"CBSS_ODS_GS":{"137":1124398818,"92":1124727398,"101":1124450643,"83":1124433278,"110":1124456872,"128":1124219743,"119":1124432547,"104":1124628336,"23":1124311329,"95":1124525375,"131":1124469343,"122":1124535024,"77":1124448871,"86":1124485192,"50":1124508471,"59":1124684544,"113":1124646095,"41":1124064203,"32":1124519454,"68":1124446354,"53":1124557152,"62":1124382089,"134":1124559698,"35":1124507617,"44":1124407028,"8":1124433001,"17":1124476420,"125":1124396390,"26":1124443966,"80":1124567565,"89":1124658207,"116":1124614654,"98":1124475544,"71":1124473950,"107":1124542529,"11":1124438204,"74":1124556643,"56":1124538763,"38":1124504126,"29":1124694681,"47":1124603677,"20":1124192614,"2":1124588467,"65":1124444154,"5":1124294139,"14":1124448849,"124":1124375546,"133":1124404400,"106":1124446976,"115":1124223326,"46":1124558944,"127":1124129305,"118":1124312050,"136":1124321917,"100":1124699504,"109":1124627319,"82":1124442521,"91":1124555678,"55":1124335963,"64":1124373113,"73":1124589828,"58":1124459124,"67":1124482459,"85":1124359891,"94":1124393984,"139":1124434665,"40":1124473917,"49":1124267140,"130":1124675189,"4":1124393400,"13":1124416361,"121":1124308740,"22":1124465366,"103":1124399814,"31":1124497397,"76":1124663123,"112":1124321464,"16":1124506270,"97":1124378643,"7":1124546058,"79":1124127427,"88":1124465497,"70":1124683417,"43":1124335991,"52":1124426670,"25":1123996945,"34":1124445672,"61":1124724327,"10":1124662549,"37":1124557737,"1":1124492770,"28":1124529006,"19":1124554560,"129":1124748745,"138":1124809837,"120":1124577963,"60":1124458993,"87":1124570910,"96":1124326628,"132":1124298235,"105":1124477087,"123":1124601206,"114":1124572178,"69":1124593835,"78":1124544523,"99":1124359214,"63":1124340260,"90":1124554065,"45":1124736498,"54":1124353713,"72":1124277878,"81":1124393509,"126":1124593551,"27":1124665644,"135":1124587297,"108":1124331277,"36":1124681226,"117":1124574196,"9":1124244797,"18":1124433233,"48":1124776095,"21":1124395649,"57":1124623562,"12":1124528086,"3":1124515899,"84":1124481011,"102":1124598175,"93":1124697719,"75":1124569607,"30":1124450004,"39":1124477532,"111":1124542936,"66":1124505569,"15":1124454816,"42":1124248608,"51":1124284882,"33":1124404806,"24":1124475933,"6":1124439966,"0":1124402328}}

batchWatermarkMs:當前event time watermark 主要是綁定遲到數(shù)據(jù)
batchTimestampMs:當前批次處理時間戳慈缔,(感覺不象unix時間藐鹤,具體有待確認)
"spark.sql.shuffle.partitions":"200" 分區(qū)數(shù)
"CBSS_ODS_GS":{"137":1124398818:topic處理的分區(qū)最新偏移量

3娱节、stateStore
存儲跨批次聚合函數(shù)運行狀態(tài)

/state/0/199/261.snapshot
--快照數(shù)據(jù),有專門的線程定期負責清理肄满,可以調(diào)整參數(shù)配置保留時間
/state/0/199/262.delta
--快照數(shù)據(jù)稠歉,有專門的線程定期負責清理蹂窖,可以調(diào)整參數(shù)配置保留時間

4、metaData
/metadata

hadoop fs -cat /metadata
{"id":"b4ead721-abef-4021-b556-1d6874ea19e3"}

最后編輯于
?著作權歸作者所有,轉載或內(nèi)容合作請聯(lián)系作者
  • 序言:七十年代末瞬测,一起剝皮案震驚了整個濱河市月趟,隨后出現(xiàn)的幾起案子恢口,更是在濱河造成了極大的恐慌,老刑警劉巖因妇,帶你破解...
    沈念sama閱讀 212,884評論 6 492
  • 序言:濱河連續(xù)發(fā)生了三起死亡事件婚被,死亡現(xiàn)場離奇詭異梳虽,居然都是意外死亡,警方通過查閱死者的電腦和手機谷炸,發(fā)現(xiàn)死者居然都...
    沈念sama閱讀 90,755評論 3 385
  • 文/潘曉璐 我一進店門拓颓,熙熙樓的掌柜王于貴愁眉苦臉地迎上來,“玉大人录粱,你說我怎么就攤上這事画拾∏嗯祝” “怎么了?”我有些...
    開封第一講書人閱讀 158,369評論 0 348
  • 文/不壞的土叔 我叫張陵适室,是天一觀的道長捣辆。 經(jīng)常有香客問我此迅,道長,這世上最難降的妖魔是什么忍些? 我笑而不...
    開封第一講書人閱讀 56,799評論 1 285
  • 正文 為了忘掉前任罢坝,我火速辦了婚禮搅窿,結果婚禮上,老公的妹妹穿的比我還像新娘男应。我一直安慰自己殉了,他們只是感情好,可當我...
    茶點故事閱讀 65,910評論 6 386
  • 文/花漫 我一把揭開白布众弓。 她就那樣靜靜地躺著,像睡著了一般脚乡。 火紅的嫁衣襯著肌膚如雪滨达。 梳的紋絲不亂的頭發(fā)上,一...
    開封第一講書人閱讀 50,096評論 1 291
  • 那天锌订,我揣著相機與錄音辆飘,去河邊找鬼蜈项。 笑死续挟,一個胖子當著我的面吹牛诗祸,可吹牛的內(nèi)容都是我干的。 我是一名探鬼主播贬媒,決...
    沈念sama閱讀 39,159評論 3 411
  • 文/蒼蘭香墨 我猛地睜開眼,長吁一口氣:“原來是場噩夢啊……” “哼漂佩!你這毒婦竟也來了投蝉?” 一聲冷哼從身側響起,我...
    開封第一講書人閱讀 37,917評論 0 268
  • 序言:老撾萬榮一對情侶失蹤关拒,失蹤者是張志新(化名)和其女友劉穎着绊,沒想到半個月后,有當?shù)厝嗽跇淞掷锇l(fā)現(xiàn)了一具尸體洲脂,經(jīng)...
    沈念sama閱讀 44,360評論 1 303
  • 正文 獨居荒郊野嶺守林人離奇死亡恐锦,尸身上長有42處帶血的膿包…… 初始之章·張勛 以下內(nèi)容為張勛視角 年9月15日...
    茶點故事閱讀 36,673評論 2 327
  • 正文 我和宋清朗相戀三年一铅,在試婚紗的時候發(fā)現(xiàn)自己被綠了潘飘。 大學時的朋友給我發(fā)了我未婚夫和他白月光在一起吃飯的照片福也。...
    茶點故事閱讀 38,814評論 1 341
  • 序言:一個原本活蹦亂跳的男人離奇死亡暴凑,死狀恐怖现喳,靈堂內(nèi)的尸體忽然破棺而出,到底是詐尸還是另有隱情犬辰,我是刑警寧澤嗦篱,帶...
    沈念sama閱讀 34,509評論 4 334
  • 正文 年R本政府宣布,位于F島的核電站幌缝,受9級特大地震影響灸促,放射性物質(zhì)發(fā)生泄漏。R本人自食惡果不足惜涵卵,卻給世界環(huán)境...
    茶點故事閱讀 40,156評論 3 317
  • 文/蒙蒙 一浴栽、第九天 我趴在偏房一處隱蔽的房頂上張望。 院中可真熱鬧轿偎,春花似錦典鸡、人聲如沸。這莊子的主人今日做“春日...
    開封第一講書人閱讀 30,882評論 0 21
  • 文/蒼蘭香墨 我抬頭看了看天上的太陽嫁乘。三九已至,卻和暖如春球碉,著一層夾襖步出監(jiān)牢的瞬間蜓斧,已是汗流浹背汁尺。 一陣腳步聲響...
    開封第一講書人閱讀 32,123評論 1 267
  • 我被黑心中介騙來泰國打工, 沒想到剛下飛機就差點兒被人妖公主榨干…… 1. 我叫王不留帮碰,地道東北人。 一個月前我還...
    沈念sama閱讀 46,641評論 2 362
  • 正文 我出身青樓斯碌,卻偏偏與公主長得像,于是被迫代替她去往敵國和親冠骄。 傳聞我的和親對象是個殘疾皇子凛辣,可洞房花燭夜當晚...
    茶點故事閱讀 43,728評論 2 351

推薦閱讀更多精彩內(nèi)容