You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@mesos.apache.org by "Gastón Kleiman (JIRA)" <ji...@apache.org> on 2019/05/08 07:06:00 UTC
[jira] [Commented] (MESOS-9513) Investigate command health check
performance
[ https://issues.apache.org/jira/browse/MESOS-9513?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16835379#comment-16835379 ]
Gastón Kleiman commented on MESOS-9513:
---------------------------------------
After grabbing a lot of perf traces of Mesos agents on DC/OS EE, the only bottleneck I could find that affects vanilla Mesos is the fact that an I/O switchboard process is launched for every DEBUG container.
If we want to improve the performance of (health) checks, we should prioritize https://issues.apache.org/jira/browse/MESOS-7121.
> Investigate command health check performance
> --------------------------------------------
>
> Key: MESOS-9513
> URL: https://issues.apache.org/jira/browse/MESOS-9513
> Project: Mesos
> Issue Type: Task
> Components: agent
> Reporter: Greg Mann
> Assignee: Gastón Kleiman
> Priority: Major
> Labels: check, foundations, health-check, mesosphere, performance
>
> Users have reported performance issues caused by too many command health checks performed too quickly by the default executor. Use of the agent's LAUNCH_NESTED_CONTAINER_SESSION call can impact the containerizer's ability to successfully launch containers in general.
> We need to investigate this issue and decide on a path forward to improve the performance of command health checks.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)