[
https://issues.apache.org/jira/browse/YUNIKORN-1213?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17631797#comment-17631797
]
Craig Condit edited comment on YUNIKORN-1213 at 11/10/22 5:27 PM:
------------------------------------------------------------------
Given the updates on YUNIKORN-1221, I would propose implementation as follows.
Since we're moving to a single ConfigMap for all of YuniKorn, we should create
a new ConfigMap entry:
{quote}{{health.checkInterval: "30s"}}
{quote}
However, we currently have no way to pass anything but the scheduler config
from the shim to the core. To allow for this, update the scheduler interface
objects responsible for configuration:
{quote}{{RegisterResourceManagerRequest {}}
{{ // ...}}
{{ // additional configuration key/value pairs}}
{{ map<string, string> extraConfig = 6;}}
{{}}}
{{UpdateConfiguration Request {}}
{{ // ...}}
{{ // additional configuration key/value pairs}}
{{ map<string, string> extraConfig = 5;}}
{{}}}
{quote}
Once these are present, we can copy the ConfigMap entries into the core on
initial registration and update (coming as part of YUNIKORN-1221). The core
should store these but ignore unknown options (since they may be for other
YuniKorn components).
I've updated the YUNIKORN-1221 design doc to show this parameter, and
YUNIKORN-1394 for the scheduler interface changes.
was (Author: ccondit):
Given the updates on YUNIKORN-1221, I would propose implementation as follows.
Since we're moving to a single ConfigMap for all of YuniKorn, we should create
a new ConfigMap entry:
{quote}{{health.checkInterval: "30s"}}
{quote}
However, we currently have no way to pass anything but the scheduler config
from the shim to the core. To allow for this, update the scheduler interface
objects responsible for configuration:
{quote}{{RegisterResourceManagerRequest {}}
{{ // ...}}
{{ // additional configuration key/value pairs}}
{{ map<string, string> extraConfig = 6;}}
{{}}}
{{UpdateConfiguration Request {}}
{{ // ...}}
{{ // additional configuration key/value pairs}}
{{ map<string, string> extraConfig = 5;}}
{{}}}
{quote}
Once these are present, we can copy the ConfigMap entries into the core on
initial registration and update (coming as part of YUNIKORN-1221). The core
should store these but ignore unknown options (since they may be for other
YuniKorn components).
I've updated the YUNIKORN-1221 design doc to show this parameter.
> The interval of the background health checker needs to be configurable
> ----------------------------------------------------------------------
>
> Key: YUNIKORN-1213
> URL: https://issues.apache.org/jira/browse/YUNIKORN-1213
> Project: Apache YuniKorn
> Issue Type: Improvement
> Components: core - scheduler
> Reporter: Weiwei Yang
> Assignee: Saad Ur Rahman
> Priority: Major
> Labels: pull-request-available
>
> YUNIKORN-1107 adds a background running health checker to verify the
> scheduler data correctness in the fixed time interval 30s:
> https://github.com/apache/yunikorn-core/blob/3ba91fb8a41c0fd0dd6243326e583dea5167199f/pkg/scheduler/health_checker.go#L34.
> We need to make this configurable, either let the user set a longer/shorter
> interval, or completely disable it.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
