## What do these changes do?
This option goes along with `min_workers`, and `max_workers`. When the
cluster is first brought up (or when it is refreshed with a subsequent
`ray up`) this number of nodes will be started.
It's a workaround for issues of scaling (see related issues) where it
can take a long time (or forever in the case where the head node has
`--num-cpus 0`) to scale up a cluster in response to increasing demand.
## Related issue number
Workaround for https://github.com/ray-project/ray/issues/3339 and https://github.com/ray-project/ray/issues/2106
* Push a warning to all users when large number of workers have been started.
* Add test.
* Fix bug.
* Give warning when worker starts instead of when worker registers.
* Fix
* Fix tests
* Limit Redis max memory to 10GB/shard by default.
* Update stress tests.
* Reorganize
* Update
* Add minimum cap size for object store and redis.
* Small test update.
This PR introduces cluster-level fault tolerance for Tune by checkpointing global state. This occurs with relatively high frequency and allows users to easily resume experiments when the cluster crashes.
Note that this PR may affect automated workflows due to auto-prompting, but this is resolvable.
## What do these changes do?
1. Fix the Jenkins test failure by add driver id to Actor GCS Key.
2. Move `object_manager_test.py` from Jenkins to Travis.
* mb impala
* fix
* paropt
* update
* cpu warn
* on cpu
* fix mb
* doc
* docs
* comment
* larger num
* early release
* remove grad clip
* only check loader count in multi gpu mode
* revert bad multigpu changes
* num sgd iter
* comment
* reuse optimizer
* add test
* par load test
* loosen test
* Update run_multi_node_tests.sh
* fix local mode
* Update agent.py
* Add a flag for whether an object has been created before
* Add regression test
* doc
* Share object directory between object and node managers
* Treat evicted actor tasks as failed
* minor
* Check return value
* Fix bug where object locations weren't getting updated on client death
* Fix mac build
* Use RayTaskError
It is possible that `test_free_objects_multi_node` would fail sometimes. If we run this test 20 times, we may found at least one failure.
The cause is that the test is based on function tasks. One raylet may create more than one worker to execute the tasks. So flush operations may be separated to several workers and not clean all the worker objects held by the plasma client.
In this PR, I change function task to actor tasks, which guarantee all the tasks are executed in one worker of a raylet.
* Add script for running stress tests.
* Add an actor tree test where actors die with some probability
* Improve test.
* Small fix
* Update tests.
* Minor change
## What do these changes do?
This passes in the right obs space to the lstm model wrapper, so that it doesn't attempt to un-flatten the already processed dict observation.
## Related issue number
Closes https://github.com/ray-project/ray/issues/3367
This PR introduces single-node fault tolerance for Tune.
## Previous behavior:
- Actors will be restarted without checking if resources are available. This can lead to problems if we lose resources.
## New behavior:
- RUNNING trials will be resumed on another node on a best effort basis (meaning they will run if resources available).
- If the cluster is saturated, RUNNING trials on that failed node will become PENDING and queued.
- During recovery, TrialSchedulers and SearchAlgorithms should receive notification of this (via `trial_runner.stop_trial`) so that they don’t wait/block for a trial that isn’t running.
Remaining questions:
- Should `last_result` be consistent during restore?
Yes; but not for earlier trials (trials that are yet to be checkpointed).
- Waiting for some PRs to merge first (#3239)
Closes#2851.
* Broadcast actor death, clean up dummy objects
* Reduce logging and clean up state when failing a task
* lint
* Make actor failure test nicer, reduce node timeout
* Suppress duplicate pre-emptive object pushes.
* Add test.
* Fix linting
* Remove timer and inline recent_pushes_ into local_objects_.
* Improve test.
* Fix
* Fix linting
* Enable retrying pull from same object manager. Randomize object manager.
* Speed up test
* Linting
* Add test.
* Minor
* Lengthen pull timeout and reissue pull every time a new object becomes available.
* Increase pull timeout in test.
* Wait for nodes to start in object manager test.
* Wait longer for nodes to start up in test.
* Small fixes.
* _submit -> _remote
* Change assert to warning.
* Make scheduling queues RemoveTasks return task states as well.
* Add test
* Don't unsubscribe for infeasible tasks when spilling over.
* Linting
* Address comments.
IMPALA support for multiagent was broken since IMPALA has a requirement that batch sizes be of a certain length. However multi-agent envs can create variable-length batches.
Fix this by adding zero-padding as needed (similar to the RNN case).
* example
* add env
* test pg
* change to test
* add atexit test
* Update rllib-env.rst
* comment
* revert unnecessary file
* fix title when actor is idle
* Update python/ray/actor.py
Co-Authored-By: ericl <ekhliang@gmail.com>
