* Remove all __future__ imports from RLlib.
* Remove (object) again from tf_run_builder.py::TFRunBuilder.
* Fix 2xLINT warnings.
* Fix broken appo_policy import (must be appo_tf_policy)
* Remove future imports from all other ray files (not just RLlib).
* Remove future imports from all other ray files (not just RLlib).
* Remove future import blocks that contain `unicode_literals` as well.
Revert appo_tf_policy.py to appo_policy.py (belongs to another PR).
* Add two empty lines before Schedule class.
* Put back __future__ imports into determine_tests_to_run.py. Fails otherwise on a py2/print related error.
* Make full use of node
implement local node
fix bugs mentioned in comments
* Add more tests
* Use more specific exception handling
* fix, lint
* fix for py2.x
* Stream logs to driver by default.
* Fix from rebase
* Redirect raylet output independently of worker output.
* Fix.
* Create redis client with services.create_redis_client.
* Suppress Redis connection error at exit.
* Remove thread_safe_client from redis.
* Shutdown driver threads in ray.shutdown().
* Add warning for too many log messages.
* Only stop threads if worker is connected.
* Only stop threads if they exist.
* Remove unnecessary try/excepts.
* Fix
* Only add new logging handler once.
* Increase timeout.
* Fix tempfile test.
* Fix logging in cluster_utils.
* Revert "Increase timeout."
This reverts commit b3846b89040bcd8e583b2e18cb513cb040e71d95.
* Retry longer when connecting to plasma store from node manager and object manager.
* Close pubsub channels to avoid leaking file descriptors.
* Limit log monitor open files to 200.
* Increase plasma connect retries.
* Add comment.
- NodeUpdater gets its' IP in parallel now (no longer in __init__)
- We use persistent connections in SSH (temp folder created only for ray; ControlMaster)
- hash_runtime_conf was performing a pointless hexlify step, wasting time on large files
- We use NodeUpdaterThreads and share the NodeProvider; NodeUpdaterProcess is removed
- AWSNodeProvider caches nodes more aggressively
- NodeProvider now has a shim batch terminate_nodes() call; AWSNodeProvider parallelises it; the autoscaler uses it
- AWSNodeProvider batches EC2 update_tags calls
- Logging changes throughout to provide standardised timing information for profiling
- Pulled out a few unnecessary is_running calls (NodeUpdater will loop waiting for SSH anyway)
## Related issue number
Issue #3599
## What do these changes do?
* distribute load and resource information on a heartbeat
* for each raylet, maintain total and available resource capacity as well as measure of current load
* this PR introduces a new notion of load, defined as a sum of all resource demand induced by queued ready tasks on the local raylet. This provides a heterogeneity-aware measure of load that supersedes legacy Ray's task count as a proxy for load.
* modify the scheduling policy to perform *capacity-based*, *load-aware*, *optimistically concurrent* resource allocation
* perform task spillover to the heartbeating node in response to a heartbeat, implementing heterogeneity-aware late-binding/work-stealing.
## What do these changes do?
1. Separate the log related code to logger.py from services.py.
2. Allow users to modify logging formatter in `ray start`.
## Related issue number
https://github.com/ray-project/ray/pull/2664
ray exec CLUSTER CMD [--screen] [--start] [--stop]
ray attach CLUSTER [--start]
Example:
ray exec sgd.yaml 'source activate tensorflow_p27 && cd ~/ray/python/ray/rllib && ./train.py --run=PPO --env=CartPole-v0' --screen --start --stop
This will in one command create a cluster and run the command on it in a screen session. The screen can later be attached to via ray attach. After the command finishes, the cluster workers will be terminated and the head node stopped.
This PR adds a driver table for the new GCS, which enables cleanup functionality associated with monitoring driver death.
Some testing in `monitor_test.py` is restored, but redis sharding for xray is needed to enable remaining tests.
* Fix documentation indentation.
* Add error table to GCS and push error messages through node manager.
* Add type to error data.
* Linting
* Fix failure_test bug.
* Linting.
* Enable one more test.
* Attempt to fix doc building.
* Restructuring
* Fixes
* More fixes.
* Move current_time_ms function into util.h.
* build_credis.sh: use an up-to-date credis commit.
* build_credis.sh: leveldb is updated, so update build cmds for it
* WIP: make monitor.py issue flush; switch gcs client to use credis
* Experimental: enable automatic GCS flushing with configurable policy.
* Fix linux compilation error
* Fix leveldb build
* Use optimized build for credis
* Address comments
* Attempt to fix tests
* Implement global state API for xray.
* Fix object table.
* Fixes for log structure.
* Implement cluster_resources.
* Add driver task to task table.
* Remove python flatbuffers code
* Get some global state API tests running.
* Python linting.
* Fix linting.
* Fix mock modules for doc
* Copy over flatbuffer bindings.
* Fix for tests.
* Linting
* Fix monitor crash.
* Add flake8 to Travis
* Add flake8-comprehensions
[flake8 plugin](https://github.com/adamchainz/flake8-comprehensions) that
checks for useless constructions.
* Use generators instead of lists where appropriate
A lot of the builtins can take in generators instead of lists.
This commit applies `flake8-comprehensions` to find them.
* Fix lint error
* Fix some string formatting
The rest can be fixed in another PR
* Fix compound literals syntax
This should probably be merged after #1963.
* dict() -> {}
* Use dict literal syntax
dict(...) -> {...}
* Rewrite nested dicts
* Fix hanging indent
* Add missing import
* Add missing quote
* fmt
* Add missing whitespace
* rm duplicate pip install
This is already installed in another file.
* Fix indent
* move `merge_dicts` into utils
* Bring up to date with `master`
* Add automatic syntax upgrade
* rm pyupgrade
In case users want to still use it on their own, the upgrade-syn.sh script was
left in the `.travis` dir.
* Use pep8 style
The original style file is actually just pep8 style, but with everything
spelled out. It's easier to use the `based_on_style` feature. Any overrides are
clearer that way.
* Improve yapf script
1. Do formatting in parallel
2. Lint RLlib
3. Use .style.yapf file
* Pull out expressions into variables
* Don't format rllib
* Don't allow splits in dicts
* Apply yapf
* Disallow single line if-statements
* Use arithmetic comparison
* Simplify checking for changed files
* Pull out expr into var
* Treat actor creation like a regular task.
* Small cleanups.
* Change semantics of actor resource handling.
* Bug fix.
* Minor linting
* Bug fix
* Fix jenkins test.
* Fix actor tests
* Some cleanups
* Bug fix
* Fix bug.
* Remove cached actor tasks when a driver is removed.
* Add more info to taskspec in global state API.
* Fix cyclic import bug in tune.
* Fix
* Fix linting.
* Fix linting.
* Don't schedule any tasks (especially actor creaiton tasks) on local schedulers with 0 CPUs.
* Bug fix.
* Add test for 0 CPU case
* Fix linting
* Address comments.
* Fix typos and add comment.
* Add assertion and fix test.
This adds (experimental) auto-scaling support for Ray clusters based on GCS load metrics. The auto-scaling algorithm is as follows:
Based on current (instantaneous) load information, we compute the approximate number of "used workers". This is based on the bottleneck resource, e.g. if 8/8 GPUs are used in a 8-node cluster but all the CPUs are idle, the number of used nodes is still counted as 8. This number can also be fractional.
We scale that number by 1 / target_utilization_fraction and round up to determine the target cluster size (subject to the max_workers constraint). The autoscaler control loop takes care of launching new nodes until the target cluster size is met.
When a node is idle for more than idle_timeout_minutes, we remove it from the cluster if that would not drop the cluster size below min_workers.
Note that we'll need to update the wheel in the example yaml file after this PR is merged.
Eric Liang