Comments (4)
oliverhu
commented on May 19, 2024
There is an API to query RM for max applicable resources, we can min(user_input, rm_max)
from tony.
erwa
commented on May 19, 2024
That would be useful. I see YARN has yarn.nodemanager.resource.cpu-vcores and yarn.nodemanager.resource.memory-mb, which seem to define the max number of vcores or memory a container can request. We could add/use something similar for GPUs.
from tony.
erwa
commented on May 19, 2024
I was testing with an older version of TonY but will try again with the latest code, which has https://github.com/linkedin/TonY/pull/12/files, which looks like it fixes this issue.
from tony.
erwa
commented on May 19, 2024
Verified that this is no longer an issue with the latest TonY code. YARN will return a container with 4 GPUs (if that's the most available on one node, even if user requested 5 GPUs), and TonY will continue with launching a TaskExecutor in the container.
from tony.
Related Issues (20)
- Task executors that support specific roles are restarted when they fail HOT 2
- Guava Conflict? HOT 2
- TonY Client allow users to specify jars to container runtime classpath HOT 2
- CI looks unstable HOT 1
- Allow that one role of task executor could make other roles exit HOT 3
- The process of task executor is still alive when existing NM marked as lost node by RM
- Configurable status when dependency times out HOT 2
- [Optimization] Seperate the interface of registerTask and getClusterSpec in TaskExecutor
- [Optimization] Introducing the config of timeout that task executor register to AM
- Get task executor's python subprocess exit detailed diagnostics message
- Instability test case of testTonyAllocationTimeoutShouldFail
- Support venv of tar.gz compression algorithm
- Introduce the simple TonY web dashboard HOT 1
- ConcurrentModificationException when we traverse registeredTasks HOT 1
- Optimize the TonY AM web dashboard page HOT 2
- Failed to get RM principal HOT 2
- Exception in thread "main" java.lang.NoClassDefFoundError: org/apache/hadoop/fs/FSDataOutputStream HOT 1
- ERROR ApplicationMaster:496 - Exception while preparing AM org.apache.hadoop.yarn.exceptions.YarnException: Can't resolve the ip of ubuntu at com.linkedin.tony.util.Utils.getHostNameOrIpFromTokenConf(Utils.java:365) at com.linkedin.tony.ApplicationMaster.prepare(ApplicationMaster.java:476) at com.linkedin.tony.ApplicationMaster.run(ApplicationMaster.java:368) at com.linkedin.tony.ApplicationMaster.main(ApplicationMaster.java:342) HOT 2
- tony-core runtime error HOT 14
- Support placement constraint
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from tony.