Ian Rodney
4160f18e46
[Autoscaler] Remove Dangerous Default Values ( #14694 )
2021-03-15 17:47:40 -07:00
Xianyang Liu
c3d8ef1b49
[autoscaler] Auto detect memory resource + P0 bug fix in k8s ( #14567 )
...
* auto detect memory resources
* format
* fixes
* fixes
* fixes
* fixes
* fixes
* address comments and fix ci
* address comments
* address comments
* corrent comments
* fixes
2021-03-16 02:32:54 +02:00
Ezra-H
eda01f9f8c
[Collective] distributed cpu tests ( #14591 )
...
Co-authored-by: Hao Zhang <zhisbug@gmail.com>
2021-03-15 16:30:57 -07:00
Edward Oakes
f1a42c4744
[serve] [ray_client] Basic test for using Ray client with Serve ( #14163 )
2021-03-15 15:30:32 -07:00
Mark
14562a2b90
Fix the docstring for ray.cancel ( #14670 )
...
* Fix the docstring for ray.cancel
* Update python/ray/worker.py
Co-authored-by: Ian Rodney <ian.rodney@gmail.com>
Co-authored-by: Ian Rodney <ian.rodney@gmail.com>
2021-03-15 15:16:17 -07:00
Edward Oakes
9f1421e3a3
[metrics] Standardize histogram to Prometheus API ( #14627 )
2021-03-15 14:37:56 -05:00
architkulkarni
8b17ec7c6d
[Serve] Disable auto conda env setting if using Ray Client ( #14672 )
2021-03-15 14:36:46 -05:00
Edward Oakes
d90cd545d1
[serve] Deprecate system-level batching with warning, update the docs ( #14648 )
2021-03-15 13:47:01 -05:00
Cassidy Laidlaw
ef5adc1476
Include py.typed in MANIFEST.in to make sure it is packaged in wheels ( #14492 )
2021-03-15 10:43:12 -07:00
Edward Oakes
c409b5b63a
[metrics] Standardize Gauge to Prometheus interface ( #14626 )
2021-03-15 12:15:26 -05:00
Ian Rodney
63d77e0d4c
[Testing] Deflake test_object_spilling ( #14678 )
2021-03-15 10:12:53 -07:00
architkulkarni
dc2be14462
[Serve] Fix bug where create_backend fails in (base) conda env ( #14633 )
2021-03-15 10:06:41 -05:00
Tao Wang
c572563e1e
[large scale]Add enable sharding option and disable sharding for gcs client ( #14600 )
2021-03-15 19:35:00 +08:00
Richard Liaw
c2aeccaf14
[tune] revert all mnist tests ( #14677 )
...
This reverts commit 3f557348a2
2021-03-14 23:58:13 -07:00
Eric Liang
3bdcca7ee5
Add basic concurrency test for ray client ( #14630 )
2021-03-13 11:24:57 -08:00
Edward Oakes
66be4801c6
Add deprecation warning to Counter.record() ( #14622 )
2021-03-12 17:04:28 -06:00
Eric Liang
b47036d014
Bump Ray client protocol version; fix dataclasses dependency for py 3.6 ( #14654 )
2021-03-12 14:58:34 -08:00
Richard Liaw
3f557348a2
[tune] re-enable MNIST tests! ( #14561 )
2021-03-12 13:35:43 -08:00
Dmitri Gekhtman
a90cffe26c
[dashboard][k8s] Better CPU reporting when running on K8s ( #14593 )
2021-03-12 12:02:15 -06:00
Raphael CHEN
c93961e070
[tune] Enable list of tuning hyperparameters in BOHB ( #14487 )
...
* [tune] Enable list of tuning hyperparameters in BOHB
* More concise code
Co-authored-by: Sumanth Ratna <sumanthratna@gmail.com>
* Add comment to `unflatten_list_dict`
* Fix lint
* Fix lint
* Add test for `unflatten_list_dict`
Co-authored-by: Sumanth Ratna <sumanthratna@gmail.com>
2021-03-12 09:22:44 -08:00
Edward Oakes
9cf328d616
[serve] Application-level batching initial commit ( #14610 )
2021-03-11 21:16:08 -06:00
Eric Liang
ee2bf0f989
Improved object store memory behavior with respect /dev/shm size ( #14629 )
2021-03-11 17:29:06 -08:00
Edward Oakes
5e2a3df7cd
Allow returning an actor handle from a remote call ( #13476 )
2021-03-11 16:52:09 -06:00
Edward Oakes
8e778d6f42
[serve] Remove more Counter.record()s ( #14628 )
2021-03-11 12:54:38 -06:00
Eric Liang
4c1df378bb
Point load_package tests to ray-project GH instead of personal ( #14605 )
2021-03-11 10:46:36 -08:00
architkulkarni
9b6d2ca345
[Core] Add runtime_env option to actor and task options, with conda_env ( #14430 )
2021-03-11 10:09:38 -06:00
Clark Zinzow
5a788474aa
[Core] First pass at privatizing non-public Python APIs. ( #14607 )
...
* async_compat
* utils
* cluster_utils
* compat
* function_manager
* import_thread
* memory_monitor
* monitor, log_monitor, ray_process_reaper
* metrics_agent
* parameter
* prometheus_exporter
* ray_logging
* signature
2021-03-10 22:47:28 -08:00
Eric Liang
081c960b59
Fix missing __init__ for wheels ( #14615 )
2021-03-10 18:13:58 -08:00
Eric Liang
4e8b53b3d0
Add an experimental load_package API ( #14552 )
2021-03-10 13:13:49 -08:00
Edward Oakes
8111ff5c3f
[serve] Use placement groups to bypass autoscaler throttling ( #13844 )
2021-03-10 13:33:44 -06:00
Edward Oakes
55a28cee52
[serve] Count -> Counter ( #14571 )
2021-03-10 11:59:44 -06:00
Eric Liang
dcb22af50d
Use vendored cloudpickle ( #14576 )
2021-03-09 22:08:45 -08:00
burglarralgrub
dfcb9c356e
Remove the --java-worker-options parameter ( #14563 )
2021-03-10 10:49:31 +08:00
Alex Wu
e1fbb8489e
[core] Supress infeasible warning ( #14068 )
2021-03-09 16:37:56 -08:00
Richard Liaw
ea7d4c6607
[placement groups] fix gpu ids for bundles ( #14574 )
2021-03-09 15:11:59 -08:00
Hao Zhang
2505bc8aa9
[Collective] Ray CPU collectives now available ( #14277 )
...
Co-authored-by: YLJALDC <dal177@ucsd.edu>
Co-authored-by: Ezra-H <huangrh9@gmail.com>
Co-authored-by: Ezra-H <44772185+Ezra-H@users.noreply.github.com>
2021-03-09 15:02:16 -08:00
Yi Cheng
ed8935406b
[core] Minimal support for runtime env ( #14270 )
2021-03-09 11:53:58 -08:00
Ian Rodney
6d5511cf80
Revert "reset memory for tasks and actors to 5% when cached memory ad…" ( #14556 )
...
This reverts commit 6f151ad510
2021-03-09 08:19:55 -08:00
Kai Fricke
43e098402a
[tune] make tune.with_parameters() work with the class API ( #14532 )
...
* [tune] make `tune.with_parameters()` work with the class API
* Update python/ray/tune/utils/trainable.py
Co-authored-by: Richard Liaw <rliaw@berkeley.edu>
Co-authored-by: Richard Liaw <rliaw@berkeley.edu>
2021-03-09 09:36:17 +01:00
Yiran Wang
a06dc39d9f
[Autoscaler] Check if SSH is available every 5 sec, not 10 ( #14484 )
2021-03-08 20:58:21 -08:00
Edward Oakes
59221b2f31
[metrics] Standardize metrics.Count API to prometheus counter ( #14498 )
2021-03-08 20:47:46 -06:00
Edward Oakes
04c009712d
Revert "Revert "Support accessing underlying attributes in RayTaskErr… ( #14449 )
2021-03-08 11:04:10 -06:00
Kai Yang
7977474899
[Core] Filter out dead nodes when getting address info from redis ( #14440 )
2021-03-08 15:48:26 +08:00
Edward Oakes
8e139046b9
[metrics] Remove unused unit field from cython classes ( #14497 )
2021-03-07 20:06:02 -06:00
Richard Liaw
dec3aa3453
Split tests for timeout ( #14516 )
2021-03-07 16:46:52 -08:00
Eric Liang
3fab5e2ada
Switch memory units to bytes ( #14433 )
2021-03-06 19:32:35 -08:00
Richard Liaw
5fc761c562
Fix test_advanced_3 timeout ( #14509 )
...
Signed-off-by: Richard Liaw <rliaw@berkeley.edu>
2021-03-06 10:59:06 -08:00
EscapeReality846089495
33b271aa97
[tune] Fixed save_to_dir w/ os.replace ( #14510 )
...
The method save_to_dir of the class Searcher in ray.tune.suggest.suggestion.py uses the os.rename method to replace tmp_search_ckpt to current ckpt. os.rename method will raise the [WinError 183] or file exists error of other operating system. os.replace is the currect way.
2021-03-06 01:14:56 -08:00
Alex Wu
2395e25fc0
[hotfix][core] Load balancing spillback feature flag ( #14457 )
2021-03-05 16:45:33 -08:00
Antoni Baum
2002cff42e
[Tune] HEBO concurrency fix after discussion with authors ( #14504 )
2021-03-05 14:05:37 -08:00
