Skip to content

Commit bcb23ab

Browse files
committed
add cloud-native docs
Signed-off-by: Jack Luar <[email protected]>
1 parent 66e402f commit bcb23ab

File tree

2 files changed

+20
-16
lines changed

2 files changed

+20
-16
lines changed

tools/AutoTuner/distributed/README.md

Lines changed: 19 additions & 15 deletions
Original file line numberDiff line numberDiff line change
@@ -23,19 +23,6 @@ For private deployments, we might have to use KubeRay
2323
2. https://www.paulsblog.dev/how-to-install-a-private-docker-container-registry-in-kubernetes/
2424
```
2525

26-
## TODO
27-
28-
- Look up how to preserve the cache during pip install.
29-
- Public flow, fixed: via autotuner script
30-
- Tune
31-
- Sweep
32-
- Public flow, fixed: via ray API.
33-
- Public flow, autoscaling
34-
- test using private registry on dockerhub same flow
35-
- Scaling concerns
36-
- increase storage of head node.
37-
- Object store memory - does that affect file transfer?
38-
3926
## Prerequisites
4027

4128
Make sure Autotuner prerequisites are installed. To do so, refer to the installation script.
@@ -44,6 +31,14 @@ Make sure Autotuner prerequisites are installed. To do so, refer to the installa
4431
pip install ray[default] google-api-python-client cryptography cloudpathlib
4532
```
4633

34+
## Cloud-native support
35+
36+
Currently, this setup supports GCP with batteries included. With configuration of the necessary variables in `public.yaml`,
37+
alternative cloud providers could be used:
38+
- [Azure](https://docs.ray.io/en/latest/cluster/vms/user-guides/launching-clusters/azure.html)
39+
- [AWS](https://docs.ray.io/en/latest/cluster/vms/user-guides/launching-clusters/aws.html)
40+
- [vSphere](https://docs.ray.io/en/latest/cluster/vms/user-guides/launching-clusters/vsphere.html)
41+
4742
## Public cluster setup
4843

4944
0a. Authenticate the necessary GCP account with enough privileges to do:
@@ -114,6 +109,15 @@ HEAD_SERVER=10.138.0.13
114109
ray job stop --address $HEAD_SERVER:6379 --no-wait {{ JOB_SUBMIT_ID }}
115110
```
116111

117-
## Private cluster setup
112+
## Roadmap
118113

119-
Coming soon.
114+
- Public flow, fixed: via autotuner script
115+
- Tune
116+
- Sweep
117+
- Public flow, fixed: via ray API.
118+
- Public flow, autoscaling
119+
- test using private registry on dockerhub same flow
120+
- Scaling concerns
121+
- increase storage of head node.
122+
- Object store memory - does that affect file transfer?
123+
- Private cluster setup

tools/AutoTuner/src/autotuner/distributed.py

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -340,7 +340,7 @@ def parse_arguments():
340340
type=str,
341341
metavar="<str>",
342342
default=None,
343-
help="Cloud storage directory for logs, defaults to None. Currently supports only GCP.",
343+
help="Cloud storage directory for logs, defaults to None. See our documentation on cloud support for more information.",
344344
)
345345

346346
# ML

0 commit comments

Comments
 (0)