Edward Oakes
5780ec1b62
Refresh ObjectIDs in raylet for stopgap GC ( #6109 )
2019-11-10 23:12:59 -08:00
Philipp Moritz
decaa65cd6
Use pickle by default for serialization ( #5978 )
2019-11-10 18:12:18 -08:00
Adam Gleave
01aee8d970
[autoscaler] Retry creating EC2 instances in new AZ ( #6129 )
2019-11-09 19:44:27 -08:00
Adam Gleave
c157e93ba1
[tune] Retry failed tasks with checkpointing disabled ( #6126 )
...
* Allow recovery for failed tasks without checkpointing
* Update docs
2019-11-09 19:35:27 -08:00
Eric Liang
afca6d3d87
Object store full with cyclic python references ( #6114 )
2019-11-08 14:08:24 -08:00
Edward Oakes
83378a8610
Improve flaky test_warning_monitor_died ( #6113 )
2019-11-08 12:11:15 -08:00
Eric Liang
4044af8520
Try to enable dashboard (again) ( #6069 )
...
* Revert "Revert "Enable the Ray dashboard by default (#5976 )" (#6068 )"
This reverts commit 1a3e97cf23
2019-11-08 10:48:48 -08:00
Eric Liang
4a28306186
Allow large returns from direct actor calls ( #6088 )
2019-11-07 21:28:55 -08:00
Edward Oakes
ca53af4d0f
Add pending task dependencies to ObjectID ref counting ( #6054 )
2019-11-07 18:37:10 -08:00
David Bignell
3f83b2daa9
[rllib] Rollout extensions ( #6065 )
...
* Rollout improvements
* Make info-saving optional, to avoid breaking change.
* Store generating ray version in checkpoint metadata
* Keep the linter happy
* Add small rollout test
* Terse.
* Update test_io.py
2019-11-05 20:34:18 -08:00
daiyaanarfeen
8f6d73a93a
[sgd] Extend distributed pytorch functionality ( #5675 )
...
* raysgd
* apply fn
* double quotes
* removed duplicate TimerStat
* removed duplicate find_free_port
* imports in pytorch_trainer
* init doc
* ray.experimental
* remove resize example
* resnet example
* cifar
* Fix up after kwargs
* data_dir and dataloader_workers args
* formatting
* loss
* init
* update code
* lint
* smoketest
* better_configs
* fix
* fix
* fix
* train_loader
* fixdocs
* ok
* ok
* fix
* fix_update
* fix
* fix
* done
* fix
* fix
* fix
* small
* lint
* fix
* fix
* fix_test
* fix
* validate
* fix
* fi
2019-11-05 11:16:46 -08:00
Mitchell Stern
82be14f943
Move gRPC calls outside of Raylet stats lock ( #6090 )
2019-11-05 00:47:15 -08:00
Philipp Moritz
fefe050a58
Fix running out of file descriptors in the WebUI ( #6086 )
2019-11-04 21:17:36 -08:00
Edward Oakes
043d1f4094
Return RayObjects to core worker ( #6052 )
2019-11-04 20:27:57 -08:00
visatish
18241f4a2d
[tune] Added resources_per_trial arg to validate_save_restore u… ( #6032 )
2019-11-04 13:24:46 -08:00
Simon Mo
c23eae5998
[Serve] Fix router-worker communication ( #5961 )
...
* Half way there, needs the strict queuing fix
* Fix scale down, use callback
* Cleanup
* Address commments
* Comment, nit
* Fix docstring
2019-11-04 11:29:21 -08:00
Eric Liang
8485304e83
Support concurrent Actor calls in Ray ( #6053 )
2019-11-04 01:14:35 -08:00
Eric Liang
fbad6f543b
Try fixing actor handle destruction on py2 ( #6076 )
2019-11-03 22:46:40 -08:00
Philipp Moritz
1c5446851a
Use Plasma with LRU refreshing integrated ( #6050 )
2019-11-03 16:19:05 -08:00
Eric Liang
1a3e97cf23
Revert "Enable the Ray dashboard by default ( #5976 )" ( #6068 )
...
This reverts commit 6166ef3e09
2019-11-01 17:08:37 -07:00
Richard Liaw
e94bebb1de
[tune] Fix Jenkins tests ( #6028 )
2019-11-01 16:42:04 -07:00
Eric Liang
fb34928a2a
[minor] Perf optimizations for direct actor task submission ( #6044 )
...
* merge optimizations
* fix
* fix memory err
* optimize
* fix tests
* fix serialization of method handles
* document weakref
* fix check
* bazel format
* disable on 2
2019-11-01 14:41:14 -07:00
Eric Liang
eef4ad3bba
Report census view data as part of raylet node stats ( #6060 )
2019-11-01 14:26:09 -07:00
Simon Mo
c8d7065bf3
[CI] Use rerunfailures instead of flaky ( #6061 )
...
* Use rerunfailures instead of flaky
* Lint
2019-11-01 13:59:03 -07:00
Eric Liang
6166ef3e09
Enable the Ray dashboard by default ( #5976 )
2019-11-01 12:19:01 -07:00
Simon Mo
7f5b3502da
Implement Detached Actor ( #6036 )
...
* Arg propagation works
* Implement persistent actor
* Add doc
* Initialize is_persistent_
* Rename persistent->detached
* Address comment
* Make test passes
* Address comment
* Python2 compatiblity
* Fix naming, py2
* Lint
2019-11-01 10:28:23 -07:00
Philipp Moritz
f7455839bf
Expose raylet info to dashboard ( #6045 )
2019-10-31 17:36:59 -07:00
Eric Liang
c86f945520
Support pass by ref args in for direct actor calls ( #6040 )
2019-10-31 16:55:10 -07:00
Edward Oakes
e9e78871b9
Remove unused function definition caching ( #6042 )
2019-10-30 16:41:18 -07:00
Simon Mo
56f3e96887
[Serve] Use ray's cloudpickle ( #6051 )
...
* Revert "Add cloudpickle as doc requirements (#6037 )"
This reverts commit 03ce3b7c5b
2019-10-30 15:21:09 -07:00
Eric Liang
8ebba202df
[minor] Reduce perf overhead of object ref tracking ( #6041 )
2019-10-29 18:14:51 -07:00
Eric Liang
b89cac976a
Basic direct actor call support in Python ( #5991 )
2019-10-28 22:09:04 -07:00
Simon Mo
9e2c5f8218
[Serve] Put global state in remote actor ( #5937 )
...
* Making progress
* Impl done, start debugging
* Tests all pass
* Add test, fix
* Update doc
* Fix type
2019-10-28 11:43:47 -07:00
Edward Oakes
c1418b04df
Remove CoreWorkerObjectInterface ( #6023 )
2019-10-28 10:48:41 -07:00
Simon Mo
e08b5d0cae
[Serve] Add a minimal cli ( #5854 )
...
* Add a minimal cli
* Integrate serve_cli with ray scripts
2019-10-28 09:51:31 -07:00
Philipp Moritz
80c01617a3
Optimize python task execution ( #6024 )
2019-10-27 00:43:34 -07:00
Stephanie Wang
eb41c945a1
Add gRPC endpoint to raylet to expose metrics ( #6005 )
2019-10-26 16:37:39 -07:00
Eric Liang
a5523466a2
Enable memstore by default ( #6003 )
2019-10-25 21:59:12 -07:00
Edward Oakes
d4055d70e3
Remove CoreWorkerTaskExecutionInterface ( #6009 )
2019-10-25 16:33:44 -07:00
Edward Oakes
e6141a0b8b
Remove UsePush logic from raylet ( #6015 )
2019-10-25 14:52:19 -07:00
Edward Oakes
f8a6ed7832
Spawn processes in background sessions ( #6008 )
...
Allows us to properly handle KeyboardInterrupts in interactive python interpreters.
2019-10-25 13:01:35 -07:00
Edward Oakes
1ce521a7f3
Remove task context from python worker ( #5987 )
...
Removes duplicated state between the python and C++ workers. Also cleans up the serialization codepaths a bit.
2019-10-25 07:38:33 -07:00
Ujval Misra
cf16b2f0c4
Add timesteps and remove ID from progress output ( #5999 )
2019-10-25 00:48:42 -07:00
Eric Liang
4edae7ea2b
Speed up task submissions a bit ( #5992 )
2019-10-25 00:10:37 -07:00
Edward Oakes
6f27d881bd
Fix core worker shutdown errors ( #6004 )
2019-10-24 22:29:05 -07:00
Edward Oakes
71a2f4c63d
fix comment ( #6006 )
2019-10-24 18:07:49 -07:00
Edward Oakes
c73fdb7425
Ignore errors in ObjectID.__dealloc__ ( #5997 )
2019-10-24 16:48:47 -07:00
Philipp Moritz
09d05bb3fa
Reduce actor submission python overhead ( #5949 )
2019-10-23 00:11:32 -07:00
Edward Oakes
02931e08f3
[core worker] Python core worker task execution ( #5783 )
...
Executes tasks via the event loop in the C++ core worker. Also properly handles signals (including KeyboardInterrupt), so ctrl-C in a python interactive shell works now (if connecting to an existing cluster).
2019-10-22 20:15:59 -07:00
Siyuan (Ryans) Zhuang
95241f6686
Fix the incorrect serialization behavior with pickle ( #5960 )
2019-10-22 18:08:36 -07:00
