Commit Graph

1519 Commits

Author SHA1 Message Date
Richard Liaw e82bea40b1 Add better analytics to docs (#1854) 2018-04-10 00:51:44 -07:00
adgirish efeaacbedc Adding support for concat (#1739)
adding tests

fixing flake8

adding init

flake 8 on test

fixing tests, imports, and flake8

handling for index

adding tests for row, index

added more robust error handling for axis

fixing test failures

cleaning up error sfor 2.7

updating travis

resolving import

fixing flake8

moved import order

Fixing to refactor and delaying implementing ray-pd inner concat

resolving ray-pd concat and from_pandas mutation

Revert "resolving ray-pd concat and from_pandas mutation"

This reverts commit 5db43e4e89e328286532f3ef98a4526575c5d08d.
2018-04-09 21:36:24 -07:00
Philipp Moritz 3039cca242 add facility to link libraries to tests (#1850) 2018-04-09 18:59:24 -07:00
Philipp Moritz 834e594709 [XRay] Register object store and raylet with the GCS (#1860) 2018-04-09 18:56:33 -07:00
Robert Nishihara 7c9e291b4b In the UI, display task breakdowns by default. (#1857) 2018-04-09 13:24:38 -07:00
Jones Wong 6b85d15bd9 supplement missed optimizer argument clip_rewards in default DQN configuration (#1853) 2018-04-08 17:58:05 -07:00
Robert Nishihara 256389dc59 Use new task spec for computing IDs in raylet code path. (#1830)
* Use new task spec for computing IDs in raylet code path.

* Fix linting.

* Fixes

* Fix test.
2018-04-08 13:31:55 -07:00
Robert Nishihara 0b7ad668ff Fix unused lambda capture compilation error. (#1844)
* Fix unused lambda capture compilation error.

* Fix linting.
2018-04-07 14:54:21 -07:00
Eric Liang 79d37ce240 [rllib] Switch to use lz4 instead of snappy (#1847)
* lz4

* comment

* updates
2018-04-07 14:43:45 -07:00
Eric Liang e6c00b2b5e [tune] Add util function to broadcast objects (#1845)
* add util

* Fri Apr  6 15:09:20 PDT 2018

* doc

* Fri Apr  6 15:21:42 PDT 2018

* Fri Apr  6 15:28:07 PDT 2018

* Fri Apr  6 15:28:26 PDT 2018

* Update tune-config.rst

* Update tune-config.rst
2018-04-07 11:37:14 -07:00
Richard Liaw bc8f62c947 [tune] Fix Median Stopping Rule Verbosity (#1833) 2018-04-06 22:58:13 -07:00
Stephanie Wang bef1d872b4 [xray] Cleanup Raylet processes on exit (#1839)
* Add raylet monitor script to timeout Raylet heartbeats

* Unit test for removing a different client from the client table

* Set node manager heartbeat according to global config

* Doc and fixes

* Add regression test for client table disconnect, refactor client table

* Convert 'Terminate' methods to destructors

* Destroy the Raylet on a SIGTERM

* Clean up workers on a SIGTERM
2018-04-06 17:21:51 -07:00
Melih Elibol 3bf80839cb Remove all runtime errors. (#1840) 2018-04-06 17:20:52 -07:00
Melih Elibol c7e11e9057 lint fix. (#1842) 2018-04-06 13:28:52 -07:00
Melih Elibol 24a8cede88 Cache object info from store notification. (#1815)
Cache all object info from object added store notification & submit to GCS via object directory.
2018-04-06 02:33:23 -07:00
Stephanie Wang bf194db4bc [xray] Basic actor support (#1835) 2018-04-06 00:17:14 -07:00
Melih Elibol 313b864e66 disconnect bug fix. (#1837) 2018-04-05 22:10:51 -07:00
Stephanie Wang cbf3181fd2 [xray] Monitor for Raylet processes (#1831)
* Add raylet monitor script to timeout Raylet heartbeats

* Unit test for removing a different client from the client table

* Set node manager heartbeat according to global config

* Doc and fixes

* Add regression test for client table disconnect, refactor client table

* Fix linting.
2018-04-05 20:45:38 -07:00
Devin Petersohn 0d9a7a3c19 [DataFrame] Update architecture to be more flexible and performant (#1821) 2018-04-05 15:14:33 -07:00
Robert Nishihara 5bde5e75e7 Implement unsafe method for flushing entire object table and task table. (#1824)
* Implement unsafe method for flushing entire object table and task table.

* Add test.

* Fix test.
2018-04-04 18:29:24 -07:00
Richard Liaw 888e70f1be [tune] HyperOpt Support (v2) (#1763) 2018-04-04 11:08:26 -07:00
Alexey Tumanov 5a9e83761d fix unused-lambda-capture on clang version 9.1 (#1823)
* fix unused-lambda-capture on clang9.1

* unused lambda capture fix continued

* lambda capture: NM

* lambda capture

* Fix linting.
2018-04-04 11:04:10 -07:00
Robert Nishihara e0193a5501 Print backtrace for RAY_LOG(FATAL) and also add file and line number … (#1805)
* Print backtrace for RAY_LOG(FATAL) and also add file and line number in common case.

* Fix linting.
2018-04-03 10:12:46 -07:00
Robert Nishihara fbfbb1c079 [xray] Integrate worker.py with raylet. (#1810)
* Integrate worker with raylet.

* Begin allowing worker to attach to cluster.

* Fix linting and documentation.

* Fix linting.

* Comment tests back in.

* Fix type of worker command.

* Remove xray python files and tests.

* Fix from rebase.

* Add test.

* Copy over raylet executable.

* Small cleanup.
2018-04-03 02:38:56 -07:00
Robert Nishihara 0fc989c6c1 Don't use 127.0.0.1 for local ip address. (#1596)
* Don't use 127.0.0.1 for ip address.

* Update test
2018-04-02 00:34:20 -07:00
Robert Nishihara d3e974a9a4 Increase ulimit -n in autoscaler examples. (#1769) 2018-04-02 00:32:56 -07:00
Robert Nishihara 27a0d58e54 Include resource string in error message for infeasible actors. (#1768) 2018-04-02 00:31:30 -07:00
Robert Nishihara 23b8793f0e Update documentation and autoscaler to find 0.4.0. (#1789) 2018-04-02 00:28:47 -07:00
Robert Nishihara 5c86f34066 Add 0.4 release blog post. (#1794) 2018-04-02 00:23:56 -07:00
Philipp Moritz 71829a2af9 [XRay] Pass in node IP address to Raylet (#1808) 2018-04-02 00:21:19 -07:00
Philipp Moritz 0bda11e009 [XRay] Fix linting (#1809) 2018-04-01 23:11:06 -07:00
Melih Elibol 6e06a9e338 XRay Task Forwarding Milestone (#1785)
Summary:
Able to run 1000 tasks with object dependencies on a set of distributed Raylets.

Raylet Changes:

Finalized ClientConnection class.
Task forwarding.
NM-to-NM heartbeats.
NM resource accounting for tasks.
Simple scheduling policy with task forwarding.
Creating and maintaining NM 2 NM long-lived connections and reusing them for task forwarding.
LineageCache Changes:

LineageCache without cleanup of tasks committed by remote nodes.
Lineage cache writeback and cleanup implementation.
ObjectManager Changes:

Object manager event loop/ClientConnection refactor.
Multithreaded object manager (disabled in this PR).
Testing Changes:

Integration tests for task submission on multiple Raylets.
Stress tests for object manager (with GCS and object store integration).


Co-authored-by: Stephanie Wang <swang@cs.berkeley.edu>
Co-authored-by: Alexey Tumanov <atumanov@gmail.com>
2018-03-31 18:02:58 -07:00
Philipp Moritz 40c9b9cd60 Fix the setuptools_scm issue (#1784) 2018-03-31 10:33:40 -07:00
Eric Liang faaa123046 [rllib] Set num_cpu=None for workers in the default settings (#1793) 2018-03-29 16:33:40 -07:00
Eric Liang 4116c64698 [tune] Remove rllib dep again, and add a test (#1792)
* tune should not depend on rllib

* fix dep test

* Tue Mar 27 16:55:41 PDT 2018

* f401
2018-03-29 15:36:49 -07:00
Stephanie Wang 925e392b2d Add an Append call to the GCS Log that checks for current length (#1788)
* TABLE_APPEND call

* Convert callbacks back to taking in a string...

* GCS returns flatbuffers, define Log class

* Cleanups

* Modify client table to use the Log interface

* Fix bug where we replied twice from redis

* Fixes

* lint

* Compile and test raylet TaskTable

* Modify GCS tables to handle unique_ptrs from nested flatbuffers

* Add raylet::TaskTable unit tests to replace ObjectTable ones

* Convert ObjectTable to a log

* Convert ObjectTable tests to the Log

* AppendAt Redis and gcs Log command

* unit test for AppendAt

* Add a Log for task reconstruction data

* Add check for unique entries in TABLE_APPEND

* Documentation
2018-03-27 13:04:43 -07:00
Robert Nishihara 8d52fe931b Add experimental feature for flushing event logs and logfiles. (#1659)
* Add experimental feature for flushing event logs and logfiles.

* Add documentation.
2018-03-27 11:57:52 -07:00
Robert Nishihara f69cbd35d4 Bump version to 0.4.0. (#1745) ray-0.4.0 2018-03-26 22:37:16 -07:00
Robert Nishihara de3cfa223d Fix monitor.py bottleneck by removing excess Redis queries. (#1786)
* Fix monitor.py bottleneck by removing excess Redis queries.

* Remove unnecessary default value.
2018-03-26 22:30:38 -07:00
Stephanie Wang 51fdbe3867 Convert the ObjectTable implementation to a Log (#1779)
* TABLE_APPEND call

* Convert callbacks back to taking in a string...

* GCS returns flatbuffers, define Log class

* Cleanups

* Modify client table to use the Log interface

* Fix bug where we replied twice from redis

* Fixes

* lint

* Compile and test raylet TaskTable

* Modify GCS tables to handle unique_ptrs from nested flatbuffers

* Add raylet::TaskTable unit tests to replace ObjectTable ones

* Convert ObjectTable to a log

* Convert ObjectTable tests to the Log
2018-03-26 20:36:48 -07:00
Robert Nishihara 1ab0d0ea69 Acquire worker lock when importing actor. (#1783) 2018-03-26 18:31:26 -07:00
Stephanie Wang 0fd4112354 Introduce a log interface for the new GCS (#1771)
* TABLE_APPEND call

* Convert callbacks back to taking in a string...

* GCS returns flatbuffers, define Log class

* Cleanups

* Modify client table to use the Log interface

* Fix bug where we replied twice from redis

* Fixes

* lint
2018-03-26 16:00:43 -07:00
Eric Liang 7c4afa4b04 [tune] Fix linting error (#1777) 2018-03-25 23:44:14 -07:00
Yan Facai (颜发才) 6b1e592d5c [tune] Added pbt with keras on cifar10 dataset example (#1729)
* [tune] Added pbt with keras on cifar10 dataset example

* ENH: add gpu resources

* CLN: requires 4 GPUs resource

* CLN: use single quotes

* CLN: don't save model by default
2018-03-25 15:57:23 -07:00
Stephanie Wang 0ad1054b8b Add a GCS table for the xray task flatbuffer (#1775)
* Introduce Task flatbuffer into xray, add to GCS

* Compile and test raylet TaskTable
2018-03-23 13:18:23 -07:00
Eric Liang 72595cca0d [tune] Change tune resource request syntax to be less confusing (#1764)
* update

* update examples

* Wed Mar 21 15:19:56 PDT 2018

* Wed Mar 21 15:21:32 PDT 2018

* Update train_a3c.py

* Update train.py

* fix resources accounting
2018-03-23 06:25:01 -07:00
Robert Nishihara 10dabce4d7 Remove from X import Y convention in RLlib ES. (#1774) 2018-03-23 05:54:31 -07:00
Christian Barra 13b3df9321 Check if the provider is external before getting the config. (#1743) 2018-03-22 22:59:29 -07:00
Stephanie Wang 8704c8618c Request and cancel notifications in the new GCS API (#1758)
* Add TableRequestNotifications and TableCancelNotifications to Redis modules

* Add RequestNotifications and CancelNotifications to generic GCS Table

* Add tests for subscribing to specific keys

* Remove TODO!

* Return the current value at the key directly from RequestNotifications instead of through publish

* Add unit test for Lookup failure callback

* Modify tests to account for empty subscription response

* Remove ObjectTable notification methods

* Clean up message parsing and doc in redis context

* Use vectors of DataT in all GCS callbacks

* Clean up SubscriptionCallback

* Move Table definitions into tables.cc

* Refactor and document redis modules

* doc

* Fix new GCS build

* Cleanups

* Revert "Fix new GCS build"

This reverts commit 6e3e69090c67ef60aaf22a9cf62be0290d989e96.

* Use vectors for internal callback interface, user-facing interface takes a reference to a single item

* Fix new GCS build

* Add unit test for Lookup failure callback

* Fix compiler errors

* Cleanup

* Publish the entry ID with the notification

* Check that the ID for a notification matches in client tests
2018-03-22 10:31:07 -07:00
Robert Nishihara 0c835a379f Fix resource bookkeeping for blocked actor methods. (#1766) 2018-03-21 20:48:04 -07:00