hiro/ray
1
0
Fork 0
mirror of https://github.com/vale981/ray synced 2025-03-09 21:06:39 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions
5519 commits 227 branches 67 tags 234 MiB
Commit graph

1419 commits

Author SHA1 Message Date
Lingxuan Zuo
7a2d7964d8
Bool config string stream must be in true or 1 (#9292) 2020-07-03 20:37:19 +08:00
ZhuSenlin
726c4cd6d4
fix actor accessor initialize inside redis gcs client (#9284) 
Co-authored-by: senlin.zsl <senlin.zsl@antfin.com>
 2020-07-03 15:48:32 +08:00
mehrdadn
29acf272b7
Build with Visual C++ (#9190) 
Co-authored-by: Mehrdad <noreply@github.com>
Co-authored-by: Simon Mo <xmo@berkeley.edu>
 2020-07-02 09:34:24 -07:00
Lingxuan Zuo
8a1cc7f8f9
Reset bool config by checking type first (#9247) 2020-07-02 18:32:41 +08:00
fangfengbin
8fcfcc4100
GCS server error handling for actor creation (#8899) 2020-07-02 16:27:32 +08:00
fangfengbin
a7a7bef622
Fix Java test crash due to incorrect destruction order (#9236) 2020-07-02 15:44:08 +08:00
Tao Wang
d35776e92d
[Core] Do not heartbeat resources if no update (#9182) 2020-07-02 13:56:49 +08:00
Lingxuan Zuo
1491508859
[Metrics] Stats supports metric exporters (#8941) 2020-07-01 10:54:55 -05:00
ChenZhilei
c11855728a
Remove raylet monitor after use GCS service (#9179) 2020-07-01 20:01:52 +08:00
chaokunyang
ab8df7afa3
Refine GetNamedActorHandle acync rpc wait (#9078) 2020-06-29 10:52:38 +08:00
Alex Wu
f8ecf6de71
Small fix (#9149) 
Fixes some basic test cases
 2020-06-28 13:43:03 -07:00
fangfengbin
1dc018d0bb
Add gcs service discovery to automatically detect GCS server restart (#8735) 2020-06-28 16:50:06 +08:00
Zhilei Chen
d78128a69c
Refactor GcsJobInfoHandle to GcsJobManager and filename (#9112) 2020-06-28 16:10:13 +08:00
mehrdadn
aac40af948
Fix some nonstandard code (#9172) 
Co-authored-by: Mehrdad <noreply@github.com>
 2020-06-28 00:22:52 -07:00
Siyuan (Ryans) Zhuang
d7549d6184
Replace arrow::Status with ray::Status in Plasma (#9154) 
* add object store status

* replace arrow status with ray status

* cleanup

* remove plasma common.cc
 2020-06-26 14:06:34 -07:00
Stephanie Wang
490cddc250
[core] Refactor distributed ref counting to remove owner task ID (#9049) 
* Add intended worker ID to GetObjectStatus, tests

* Remove TaskID owner_id

* lint

* Update message

* lint

* Fix build
 2020-06-25 17:55:03 -07:00
mehrdadn
ab3413c124
Make process termination more robust (#9081) 2020-06-25 15:30:53 -05:00
Eric Liang
4522038259
[iter] Add .transform() function for arbitrary generator transforms (#8978) 2020-06-25 11:04:14 -07:00
Siyuan (Ryans) Zhuang
80bcbe20c7
[Core] Remove object notification IPC between Plasma and Raylet (initial step) (#8939) 
* initial refactoring

redirect notifications to eventloop

implement direct notifications

* protect vector with mutex
 2020-06-24 13:54:40 -07:00
Siyuan (Ryans) Zhuang
613abdf1b6
Remove arrow macros in plasma store (#9115) 2020-06-23 23:34:44 -07:00
Siyuan (Ryans) Zhuang
acb7270bd7
Adopt upstream plasma changes (#9061) 
* adopt upstream plasma changes
 2020-06-23 14:19:57 -07:00
Simon Mo
b6d425526d
Move actor task submission to io service (#9093) 2020-06-23 10:07:33 -07:00
Siyuan (Ryans) Zhuang
306ca75737
Fix ray arrow logs (#9097) 
* convert arrow logs to ray logs

* remove extra plasma tests and modules
 2020-06-23 10:02:30 -07:00
Zhilei Chen
8f2564f1a6
fix a bug that move a const variable (#9080) 2020-06-23 11:54:18 +08:00
Siyuan (Ryans) Zhuang
7a110b9401
[Core] Remove digests in plasma (4x performance improvement) (#8980) 
* remove digest in plasma

* totally remove list
 2020-06-22 14:24:32 -07:00
mehrdadn
275da2e400
Fix Google log directory again (#9063) 2020-06-22 14:56:28 -05:00
mehrdadn
1a40d24174
Handle loop_ NULL case (#9067) 
Co-authored-by: Mehrdad <noreply@github.com>
 2020-06-22 11:05:29 -07:00
SangBin Cho
e254dd3115
Do not add reference count when it is local mode. (#8979) 2020-06-21 16:01:06 -05:00
mehrdadn
981f67bfb0
Fix more Windows issues (#9011) 
Co-authored-by: Mehrdad <noreply@github.com>
 2020-06-19 18:51:45 -07:00
mehrdadn
f8d49d69c1
Fix and merge asio client read/write operations (#9026) 
Co-authored-by: Mehrdad <noreply@github.com>
 2020-06-19 18:49:55 -07:00
Gabriele Oliaro
311c55132c
redefined SchedulingClass to avoid including the FunctionDescriptor (#9022) 
* redefined SchedulingClass to avoid including the FunctionDescriptor

* updated TestSchedulingKeys test in DirectTaskTransportTest
 2020-06-19 13:12:48 -07:00
mehrdadn
f43cad6371
Fix Google log directory (#9025) 2020-06-19 11:00:02 -05:00
Zhilei Chen
d8a9247448
Remove gcs_service_disabled ci jobs and code (#8854) 2020-06-19 11:32:27 +08:00
SangBin Cho
110f88ff61
Improve raylet failure msg. (#8986) 
* Use better error messages.

* Improve message.

* Fix message based on code review.
 2020-06-18 13:36:32 -07:00
Siyuan (Ryans) Zhuang
4654b3c07a
Use raylet signal handling in plasma store when running plasma store as a thread (#9007) 2020-06-18 11:52:06 -07:00
Edward Oakes
8a99fd205e
[dashboard] Pipe resource assignments to dashboard (#8998) 2020-06-18 11:14:59 -05:00
Zhilei Chen
0de2efd330
Fix a bug that 'remote_avaiable' used after it was moved #9002(#9002) 2020-06-18 18:04:42 +08:00
fangfengbin
c295284370
Optimize gcs server resubscribe (#8896) 2020-06-17 20:05:50 +08:00
Tao Wang
9f0f542660
Remove actor table info from storage when a driver exits (#8761) 
* delete contents of table related to specified job when the job is dead

* check status

* implement GetByJobId in gcs table storage

* add test case

* add test case

* fix test cases

* expose MGET and make match_pattern only related with SCAN

* add test case for table storage

* delete checkpoint

* make MGetValues static

* add most test case

* add object test case

* avoid accessing to storage when get matched object ids per job id

* rename job info handler

* use listener to sense job finished

* clear actor state

* add comments, remove actions in task handler

* let raylet do object cleaning. only remove non-detached actors

* only remove informations of non-detached actor

* remove unused methods
 2020-06-16 18:43:08 -07:00
Stephanie Wang
fa16c7666a
Fix possible deadlock in CoreWorkerDirectActorTaskSubmitter (#8973) 2020-06-16 15:30:15 -07:00
fangfengbin
4facac023f
Fix heap-use-after-free bug of gcs pub sub testcase (#8968) 2020-06-16 21:00:37 +08:00
Siyuan (Ryans) Zhuang
b68fede30b
Convert include guard to pragma once (#8957) 2020-06-16 01:29:43 -07:00
mehrdadn
101c215125
Get more tests running on Windows (#6537) 
* Get rid of system() calls

* Work around '/usr/share/mini' showing up on GitHub Actions (probably due to psutil truncation)

https://github.com/ray-project/ray/runs/722480047?check_suite_focus=true

* Don't check for socket max path length on Windows

* Don't check for socket existence on Windows

* Fix race condition in Windows fate-sharing

* Work around missing .exe extension for Redis tests

* Add more tests to GitHub Actions

Co-authored-by: Mehrdad <noreply@github.com>
 2020-06-12 21:32:10 -07:00
Siyuan (Ryans) Zhuang
ed77c8b16c
[Core] Use global variable to eliminate force thread termination in plasma (#8912) 
* use global variable to eliminate force thread termination
 2020-06-12 14:20:53 -07:00
Siyuan (Ryans) Zhuang
4b31b383f3
[Core] Run Plasma Store as a Raylet thread (with a feature flag) (#8897) 
* integrate plasma store as a thread (C++)

* integrate plasma store as a thread (Python)

* fix config issues

* remove plasma component fail tests

* without forcefully kill the plasma store thread
 2020-06-11 22:54:08 -07:00
mehrdadn
cae475c46a
Fix Windows build (#8905) 
Co-authored-by: Mehrdad <noreply@github.com>
 2020-06-11 14:54:37 -07:00
Stephanie Wang
05010caed2
[core] Fix race condition for object reconstruction (#8791) 
* Fix

* doc

* Unit test

* Update src/ray/core_worker/task_manager.h

Co-authored-by: Edward Oakes <ed.nmi.oakes@gmail.com>

* Update src/ray/core_worker/task_manager.h

Co-authored-by: Edward Oakes <ed.nmi.oakes@gmail.com>

* Update src/ray/core_worker/task_manager.h

Co-authored-by: Edward Oakes <ed.nmi.oakes@gmail.com>

* lint

Co-authored-by: Edward Oakes <ed.nmi.oakes@gmail.com>
 2020-06-10 19:49:12 -07:00
Ian Rodney
2cf3d8c92c
[core] Check that port is unused before assigning to worker (#8773) 2020-06-10 18:35:38 -05:00
fangfengbin
a5bebd4408
Fix create actor rpc reconnect bug (#8855) 2020-06-10 10:53:53 +08:00
Siyuan (Ryans) Zhuang
3d473600a8
[Core] Use Ray ObjectID in Plasma (#8852) 
* Use Ray ObjectIDs instead

* remove unused code
 2020-06-09 10:10:49 -07:00