Commit graph

7852 commits

Author SHA1 Message Date
Edward Oakes
2274a94453
[serve] Remove deprecated system-level batching (#15065) 2021-04-06 17:58:42 -05:00
Yi Cheng
35b45a03cf
[runtime_env] Add working_directory into sys.path (#15145) 2021-04-06 17:49:06 -05:00
Alex Wu
10fdb9e9ac
[metrics] Scheduler metrics (#14716) 2021-04-06 11:27:54 -07:00
Siyuan (Ryans) Zhuang
64cc092959
[Core] Cleanup C++ code (#15109)
* cleanup c++ code

* more cleanup

* lint

* lint
2021-04-06 03:29:03 -07:00
Stefan Schneider
49ba51979e
Functions for restoring from last or best checkpoint (#14735)
Adds a helper function to retrieve the latest checkpoint after selecting the best trial according to a metric.
2021-04-06 12:19:09 +02:00
Edward Oakes
43932f5330
[serve] Health checking in the controller instead of using max_restarts (#15047) 2021-04-05 21:22:40 -05:00
Yi Cheng
e552e3f19c
Skip test_dependency_torch (#15123) 2021-04-05 18:02:10 -07:00
Clark Zinzow
f17e59093d
[dask-on-ray] Fix Dask-on-Ray scheduler break caused by changing internal Dask API. (#15100) 2021-04-05 17:58:44 -07:00
Yi Cheng
afc92130fa
Unpack runtime env to runtime_resource (#15111) 2021-04-05 17:35:31 -07:00
Dmitri Gekhtman
410f768046
[Kubernetes] [Dashboard] Remove disk data from dashboard when running on K8s. (#14676) 2021-04-05 17:16:20 -07:00
Siyuan (Ryans) Zhuang
6f56d7e360
Fix compilation warnings (#15104) 2021-04-05 16:14:32 -07:00
Dmitri Gekhtman
ee291ae762
[kubernetes][test][minor] Fix K8s test by an adding an underscore. (#15122) 2021-04-05 15:18:20 -07:00
Ian Rodney
29566c3527
[Code Cleanup] Rename to Allowlist + Denylist (#15117)
Co-authored-by: Simon Mo <simon.mo@hey.com>
2021-04-05 13:58:18 -07:00
Ian Rodney
c1bffc2784
[Core] Pick Up GCS Port From Env Variable (#15097) 2021-04-05 13:57:55 -07:00
Richard Liaw
6269140ef8
[minor] improve warning message for Ray (#15005) 2021-04-05 12:39:42 -07:00
Yi Cheng
5806e726f4
[core] Internal kv support in gcs (#14656)
* server side ready

* client size

* py

* fix

* up

* format

* add files

* add pyx

* up

* up

* up

* add keys

* format

* update

* format

* add unittests

* add files

* up

* up

* fix

* up

* fix thread issue

* format

* fix

* Fix

* format

* fix

* more

* fix conflict

* fix

* fix order

* format

* compiling fix

* lint

* fix

* fix some

* some fix

* fix comment

* fix name

* format

* fix compatible issue

* fix name

* fix lint

* disconnect safe

* up

* format

* fix

Co-authored-by: Yi Cheng <singye888@gmail.com>
2021-04-05 10:26:46 -07:00
Richard Liaw
e72f6b0377
Fix ray[full] -> ray[cluster] #15112
Signed-off-by: Richard Liaw <rliaw@berkeley.edu>
2021-04-05 09:55:00 -07:00
valiantljk-tt
9b85b0b82b
[runtime_env] Continue hashing if data is none (#14906) 2021-04-05 11:24:09 -05:00
architkulkarni
0d1b30620b
[Serve] Un-revert #15000 Set controller and HTTP proxy num_cpus=0 by default (#15092) 2021-04-05 09:58:55 -05:00
Dmitri Gekhtman
47a3bd7b67
[kubernetes][autoscaler] Rename autoscaler K8s module (#15101) 2021-04-04 15:30:00 -07:00
Yi Cheng
672dad8056
Fix gcs test failure (#15098) 2021-04-04 14:53:04 +08:00
Dmitri Gekhtman
e0ed549006
return False, don't raise (#15107) 2021-04-03 14:04:50 -07:00
Yi Cheng
1920db0a64
Deflak dask related task on win32 (#15102)
* disable dask test on win32

* disable dask test on win32

* disable dask test on win32

* disable dask test on win32
2021-04-03 12:26:29 -07:00
Richard Liaw
e9f365ad2e
[tune] support skopt/bayesopt search restore (#15075) 2021-04-03 12:24:30 -07:00
Siyuan (Ryans) Zhuang
7fd86f7e15
[Core] Use static callback instead of dynamic notification listener (#15059)
* static callback & remove outdated protocol

* address comments

* fix

* make fields constant

* fix windows compilation error
2021-04-02 22:33:41 -07:00
SangBin Cho
cef6286f63
[Pubsub] Batch messages (#15084)
* batch pubsub 1

* Logic done. Tests left.

* done.
2021-04-02 16:42:18 -07:00
Alex Wu
aea28c53ce
. (#15093)
Co-authored-by: Alex <alex@anyscale.com>
2021-04-02 16:39:38 -07:00
Yi Cheng
4caf7a511d
Deflaky test failure in win32 (#15090)
* deflake win32 failure

* skip dask on w32
2021-04-02 14:56:24 -07:00
SangBin Cho
015369db34
[Core] Fix plasma store segfault (#15071)
* Use shared pointer instead of a raw pointer

* Lint.

* Addressed code review.

* Addressed code review.g
2021-04-02 14:54:20 -07:00
architkulkarni
d8f8583e80
Revert "[Serve] Set controller and HTTP proxy num_cpus=0 by default (#15000)" (#15091)
This reverts commit 39aa01fc2c.
2021-04-02 13:01:57 -07:00
Yi Cheng
ecb94b3fe9
Add test case to check job conf compatible issue (#15082) 2021-04-02 12:03:21 -07:00
Dmitri Gekhtman
42565d5bbe
[autoscaler] Fix update/terminate race condition (#15019)
Co-authored-by: AmeerHajAli <ameerh@berkeley.edu>
2021-04-02 11:57:02 -07:00
SangBin Cho
3578d4e9d8
[Object Spilling] Limit number of objects to fuse (#15034)
* ready to go.

* Done.

* done.

* Done.

* Addressed code review.

* Fix a build issue.
2021-04-02 10:49:15 -07:00
Edward Oakes
96cc7897f7
[serve] Use longest prefix matching for path routing (#15041) 2021-04-02 12:01:47 -05:00
architkulkarni
39aa01fc2c
[Serve] Set controller and HTTP proxy num_cpus=0 by default (#15000) 2021-04-02 12:01:22 -05:00
Dmitri Gekhtman
6f81ec1998
[kubernetes][test] Operator test tweaks. (#15074) 2021-04-02 09:20:52 -07:00
Kai Fricke
8de66fce3d
[tune] Improve BOHB/ConfigSpace dependency check (#15064) 2021-04-02 10:19:49 +02:00
SangBin Cho
3965310f93
[Core] Fix the check failure from object manager (#15070) 2021-04-01 21:21:42 -07:00
Alex Wu
f52c855704
[core] Fix placement group GPU assignment bug (#15049) 2021-04-01 17:46:09 -07:00
Yi Cheng
d4c20c970b
[core] Fix UTIL worker issue (#14925)
* Fix

* format

* more

* format

* fix

* fix

* fix comment

* fix test failure
2021-04-01 17:36:45 -07:00
Simon Mo
c9dac9328e
[Serve] Fix serializing nested fields in Pydantic (#15069) 2021-04-01 17:20:34 -07:00
Siyuan (Ryans) Zhuang
6ad379864e
[doc] Fix inconsistent doc about ObjectID bytes (#15072) 2021-04-01 17:14:30 -07:00
Alex Wu
4fba05ae4d
[core] Hybrid scheduling policy. (#14790) 2021-04-01 16:59:59 -07:00
Dmitri Gekhtman
474fb6bf0c
[kubernetes][client][docs] Note requirement for matching Ray versions (#15068) 2021-04-01 15:08:25 -07:00
Edward Oakes
346994745a
[serve] Get handle in starlette endpoint constructor instead of lazily (#15066) 2021-04-01 16:07:28 -05:00
Ian Rodney
22c1aeb240
[Tests] Skip autoscaler tests on Windows (#15033) 2021-04-01 10:16:42 -07:00
fangfengbin
18728b2b7e
Fix c++ gcs test bug (#15063)
* fix ut bug

* fix bug

Co-authored-by: 灵洵 <fengbin.ffb@antgroup.com>
2021-04-01 09:19:24 -07:00
SangBin Cho
005cff0092
Revert "Revert "[Core] Implement long polling-based pubsub to reduce … (#14909) 2021-04-01 09:03:15 -07:00
Eric Liang
b2c5093054
Disable flaky windows://python/ray/tests:test_gcs_fault_tolerance (#15052) 2021-04-01 10:54:43 -05:00
Tomas Babej
bc42e69503
ci: Fix broken symlink detection (#15054) 2021-04-01 08:33:51 -07:00