ReleaseBaidu (ERNIE)Baidu (ERNIE)published Mar 23, 2023seen 5d

PaddlePaddle/PaddleFlow v0.14.6

PaddlePaddle/PaddleFlow

Open original ↗

Captured source

source ↗
published Mar 23, 2023seen 5dcaptured 9hhttp 200method plain

v0.14.6 🌈

Repository: PaddlePaddle/PaddleFlow

Tag: v0.14.6

Published: 2023-03-23T07:37:49Z

Prerelease: no

Release notes:

Exciting Changes

  • s3 api not allow to delete over 1k obj at once, rename op may trigger… @caizejun (#1081)
  • CFS Mkdir perm @luoyuedong (#1085)
  • fix: mount pod cpu and memory limit bug @luoyuedong (#1082)
  • close dir do not need to reduce filehandler count @caizejun (#1079)
  • fix: sts duration cause panic bug @luoyuedong (#1077)
  • fix: minio list bug @luoyuedong (#1074)
  • fix: sts bug @luoyuedong (#1073)
  • fs client remove all have not check if a entry is dir @caizejun (#1064)
  • fixbug: support init k3s, and fix create job create ns failed @qiaoshuangshuang (#1070)
  • paddleflow-deployment.yaml 1.4.5.2,Add operation - n paddleflow @liujiaxing123666 (#1067)
  • [pipeline]: bugfix: no resume for run @HaozhengAN (#1055)
  • fix: sts delete bug @luoyuedong (#1060)
  • Revert "fs client remove all have not check if a entry is dir" @hysunflower (#1061)
  • fs client remove all have not check if a entry is dir @caizejun (#1059)
  • fix: The name of pv is inconsistent when creating jobs concurrently @luoyuedong (#1058)
  • bos can not get sts with properity sts not equal true @caizejun (#1052)
  • feat(job): adjust log url format @D0m021ng (#1050)
  • perf: optimize s3 stat and write @luoyuedong (#1049)
  • feature: all filesysetem support chown and chmod @luoyuedong (#1048)
  • feature: all filesystem support chown and chmod @luoyuedong (#1046)
  • feat(job): generate log url for paddleflow job @D0m021ng (#1033)
  • pread use too many mem @caizejun (#1042)
  • [pipeline] cherry-pick from 145: fatal error: concurrent map writes (#1026) @HaozhengAN (#1035)
  • feature: hdfs chown support @luoyuedong (#1038)
  • simplify afs mount ins @caizejun (#1039)
  • [pipeline] add metrics for pipeline, run and schedule @HaozhengAN (#1015)
  • fix: csi secret env @luoyuedong (#1034)
  • fix: csi mount pod need add secret key @luoyuedong (#1032)
  • feature: add secret key in paddleflow and csi support csi @luoyuedong (#1031)
  • feature: add bos sts type @luoyuedong (#1016)
  • fix warmup bug @caizejun (#1030)
  • feat(cluster resources): support list cluster resources by queue @D0m021ng (#1011)
  • Test op @liujiaxing123666 (#1029)
  • fix: cache path mont propagation @luoyuedong (#1022)
  • feature: add fs cache go-sdk @luoyuedong (#1014)
  • feature: add 1.13 csi-plugin install docs @luoyuedong (#1013)
  • feature: add sts paddleflow server @luoyuedong (#1008)
  • [pipeline] opt return code and return value of some api according to testi result @HaozhengAN (#1010)
  • feature: add k3s runtime, and only support single job @qiaoshuangshuang (#987)
  • add ns check and create if not exist @tornado404 (#1005)
  • feat(job): support aware of request resources is changed while pod is running @D0m021ng (#1006)
  • feature: add afs mount @luoyuedong (#1004)
  • feat(monitor): add grafana dashboard for volcano @D0m021ng (#1003)
  • Revert "Revert "reduce dockerfile and support muti-arch binary and image"" @tornado404 (#1001)
  • fix: hdfs abcExecption bug fix @luoyuedong (#998)
  • Revert "reduce dockerfile and support muti-arch binary and image" @tornado404 (#1000)
  • reduce dockerfile and support muti-arch binary and image @tornado404 (#986)
  • [pieline] bugfix: fix bug that root user cannot user the fs which belong to other user @HaozhengAN (#996)
  • fix: cfs mount bug @luoyuedong (#995)
  • [pipeline] optimize the msg of Run @HaozhengAN (#993)
  • feature(job): add patch for custom single job @D0m021ng (#992)
  • fix: writer close bug @luoyuedong (#989)
  • [pipeline]: Optimize interface return value @HaozhengAN (#975)
  • feature: cfs mount @luoyuedong (#988)
  • fix write not flush, reader read emtpy and meta size not zero bug @luoyuedong (#985)
  • reduce relations between job and task Listener @tornado404 (#983)
  • security fix, fix python cli pyyaml from 5.3 to 5.4 @qiaoshuangshuang (#984)
  • fix reader read err maybe cause dead read loop @luoyuedong (#980)
  • improve job handler by dynamic watch @tornado404 (#981)
  • delete accesskey @caizejun (#978)
  • fix stop job failed @tornado404 (#976)
  • fix append error @luoyuedong (#973)
  • fix autopatch resource @tornado404 (#970)
  • IsSkipResourceValidate @tornado404 (#971)
  • Fix vc queue weight is empty @D0m021ng (#968)
  • [pipeline] fix bug: there is no endTime where job is terminated @HaozhengAN (#966)
  • feat(job): refactor runtime job struct @D0m021ng (#965)
  • support env limit @tornado404 (#959)
  • feat(ut): remove useless spark apis code @D0m021ng (#963)
  • feat(job): adjust job event handler code @D0m021ng (#962)
  • delete file @luoyuedong (#961)
  • fix: hdfs open and create maybe conflict @luoyuedong (#958)
  • feature: add mount warmup tool and increase gc interval @luoyuedong (#954)
  • support using default queue @tornado404 (#956)
  • bug fix: can't read too many direntrys @caizejun (#953)
  • Fix: return alloctable gpu resoruces with uniform format @D0m021ng (#952)
  • fix: hdfs append bug @luoyuedong (#948)
  • fix log range @tornado404 (#949)
  • add log client @tornado404 (#947)
  • add paddle job resource support when using extensiontemplate @tornado404 (#944)

What's Changed

  • add paddle job resource support when using extensiontemplate by @tornado404 in https://github.com/PaddlePaddle/PaddleFlow/pull/944
  • add log client by @tornado404 in https://github.com/PaddlePaddle/PaddleFlow/pull/947
  • fix log range by @tornado404 in https://github.com/PaddlePaddle/PaddleFlow/pull/949
  • fix: hdfs append bug by @luoyuedong in https://github.com/PaddlePaddle/PaddleFlow/pull/948
  • Fix: return alloctable gpu resoruces with uniform format by @D0m021ng in https://github.com/PaddlePaddle/PaddleFlow/pull/952
  • bug fix: can't read too many direntrys by @caizejun in https://github.com/PaddlePaddle/PaddleFlow/pull/953
  • support using default queue by @tornado404 in https://github.com/PaddlePaddle/PaddleFlow/pull/956
  • feature: add mount warmup tool and increase gc interval by @luoyuedong in https://github.com/PaddlePaddle/PaddleFlow/pull/954
  • fix: hdfs open and create maybe conflict by @luoyuedong in https://github.com/PaddlePaddle/PaddleFlow/pull/958
  • delete file by @luoyuedong in https://github.com/PaddlePaddle/PaddleFlow/pull/961
  • feat(job): adjust job event handler code by @D0m021ng in https://github.com/PaddlePaddle/PaddleFlow/pull/962
  • feat(ut): remove useless spark apis code by @D0m021ng in https://github.com/PaddlePaddle/PaddleFlow/pull/963
  • support env limit by @tornado404 in…

Excerpt shown — open the source for the full document.