You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@yunikorn.apache.org by "Peter Bacsko (Jira)" <ji...@apache.org> on 2024/01/05 15:18:00 UTC
[jira] [Updated] (YUNIKORN-2308) Mock scheduler performance test improvements
[ https://issues.apache.org/jira/browse/YUNIKORN-2308?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Peter Bacsko updated YUNIKORN-2308:
-----------------------------------
Description:
In YUNIKORN-1748, we added a test case which starts Yunikorn with a mock K8s environment and performs a throughput measurement.
There are ways to improve this:
# The pods are not updated to Running/Completed state. To have a more realistic test, let's add a logic which keeps updating bound pods.
# The configuration is very simple: 6 leaf queues directly under root. We don't have labels, affinities, selectors, queue or user/group quotas. Add a new test case which measures performance when using a more complicated setup.
was:
In YUNIKORN-1748, we added a test case which starts Yunikorn with mock K8s environment and performs a throughput measurement.
There are ways to improve this:
# The pods are not updated to Running/Completed state. To have a more realistic test, let's add a logic which keeps updating bound pods.
# The configuration is very simple: 6 leaf queues directly under root. We don't have labels, affinities, selectors, queue or user/group quotas. Add a new test case which measures performance when using a more complicated setup.
> Mock scheduler performance test improvements
> --------------------------------------------
>
> Key: YUNIKORN-2308
> URL: https://issues.apache.org/jira/browse/YUNIKORN-2308
> Project: Apache YuniKorn
> Issue Type: Improvement
> Components: shim - kubernetes, test - smoke
> Reporter: Peter Bacsko
> Assignee: Peter Bacsko
> Priority: Minor
>
> In YUNIKORN-1748, we added a test case which starts Yunikorn with a mock K8s environment and performs a throughput measurement.
> There are ways to improve this:
> # The pods are not updated to Running/Completed state. To have a more realistic test, let's add a logic which keeps updating bound pods.
> # The configuration is very simple: 6 leaf queues directly under root. We don't have labels, affinities, selectors, queue or user/group quotas. Add a new test case which measures performance when using a more complicated setup.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@yunikorn.apache.org
For additional commands, e-mail: issues-help@yunikorn.apache.org