[jira] [Commented] (FLINK-25648) Redundant to query deployment when creating task manager pods
[ https://issues.apache.org/jira/browse/FLINK-25648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17507370#comment-17507370 ] Yang Wang commented on FLINK-25648: --- [~Yu.an] Thanks for your contribution. > Redundant to query deployment when creating task manager pods > - > > Key: FLINK-25648 > URL: https://issues.apache.org/jira/browse/FLINK-25648 > Project: Flink > Issue Type: Improvement > Components: Deployment / Kubernetes >Affects Versions: 1.14.2 >Reporter: Yuan Huang >Assignee: Yuan Huang >Priority: Critical > Labels: pull-request-available > Attachments: trace.png > > > When creating Task Manager (TM) Pod, the fabric client needs the deployment > to set owner reference for the TM pod. However, repeating to querying the > deployment each time is unnecessary which will waste a lot of time. > [https://github.com/apache/flink/blob/36ff71f5ff63a140acc634dd1d98b2bb47a76ba5/flink-kubernetes/src/main/java/org/apache/flink/kubernetes/kubeclient/Fabric8FlinkKubeClient.java] > (line 128-162) > > According to the measured result by > [https://arthas.aliyun.com/doc/trace.html#id1]) below: Quering deployment may > cost more than 50% (203/349 ms) of time. > Can we save the deployment reference in memory instead of querying it each > time to save pod creating time? > !trace.png! -- This message was sent by Atlassian Jira (v8.20.1#820001)
[jira] [Commented] (FLINK-25648) Redundant to query deployment when creating task manager pods
[ https://issues.apache.org/jira/browse/FLINK-25648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17505957#comment-17505957 ] Yuan Huang commented on FLINK-25648: - Hi, [~wangyang0918]. Thank you for your advice. I updated the code. > Redundant to query deployment when creating task manager pods > - > > Key: FLINK-25648 > URL: https://issues.apache.org/jira/browse/FLINK-25648 > Project: Flink > Issue Type: Improvement > Components: Deployment / Kubernetes >Affects Versions: 1.14.2 >Reporter: Yuan Huang >Assignee: Yuan Huang >Priority: Critical > Labels: pull-request-available > Attachments: trace.png > > > When creating Task Manager (TM) Pod, the fabric client needs the deployment > to set owner reference for the TM pod. However, repeating to querying the > deployment each time is unnecessary which will waste a lot of time. > [https://github.com/apache/flink/blob/36ff71f5ff63a140acc634dd1d98b2bb47a76ba5/flink-kubernetes/src/main/java/org/apache/flink/kubernetes/kubeclient/Fabric8FlinkKubeClient.java] > (line 128-162) > > According to the measured result by > [https://arthas.aliyun.com/doc/trace.html#id1]) below: Quering deployment may > cost more than 50% (203/349 ms) of time. > Can we save the deployment reference in memory instead of querying it each > time to save pod creating time? > !trace.png! -- This message was sent by Atlassian Jira (v8.20.1#820001)
[jira] [Commented] (FLINK-25648) Redundant to query deployment when creating task manager pods
[ https://issues.apache.org/jira/browse/FLINK-25648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17504701#comment-17504701 ] Yang Wang commented on FLINK-25648: --- Sorry for the late response. I have left two comments in the PR. Please have a look. > Redundant to query deployment when creating task manager pods > - > > Key: FLINK-25648 > URL: https://issues.apache.org/jira/browse/FLINK-25648 > Project: Flink > Issue Type: Improvement > Components: Deployment / Kubernetes >Affects Versions: 1.14.2 >Reporter: Yuan Huang >Assignee: Yuan Huang >Priority: Critical > Labels: pull-request-available > Attachments: trace.png > > > When creating Task Manager (TM) Pod, the fabric client needs the deployment > to set owner reference for the TM pod. However, repeating to querying the > deployment each time is unnecessary which will waste a lot of time. > [https://github.com/apache/flink/blob/36ff71f5ff63a140acc634dd1d98b2bb47a76ba5/flink-kubernetes/src/main/java/org/apache/flink/kubernetes/kubeclient/Fabric8FlinkKubeClient.java] > (line 128-162) > > According to the measured result by > [https://arthas.aliyun.com/doc/trace.html#id1]) below: Quering deployment may > cost more than 50% (203/349 ms) of time. > Can we save the deployment reference in memory instead of querying it each > time to save pod creating time? > !trace.png! -- This message was sent by Atlassian Jira (v8.20.1#820001)
[jira] [Commented] (FLINK-25648) Redundant to query deployment when creating task manager pods
[ https://issues.apache.org/jira/browse/FLINK-25648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17504155#comment-17504155 ] Yuan Huang commented on FLINK-25648: - Hi [~wangyang0918], just a reminder, is there any update for this PR? > Redundant to query deployment when creating task manager pods > - > > Key: FLINK-25648 > URL: https://issues.apache.org/jira/browse/FLINK-25648 > Project: Flink > Issue Type: Improvement > Components: Deployment / Kubernetes >Affects Versions: 1.14.2 >Reporter: Yuan Huang >Assignee: Yuan Huang >Priority: Critical > Labels: pull-request-available > Attachments: trace.png > > > When creating Task Manager (TM) Pod, the fabric client needs the deployment > to set owner reference for the TM pod. However, repeating to querying the > deployment each time is unnecessary which will waste a lot of time. > [https://github.com/apache/flink/blob/36ff71f5ff63a140acc634dd1d98b2bb47a76ba5/flink-kubernetes/src/main/java/org/apache/flink/kubernetes/kubeclient/Fabric8FlinkKubeClient.java] > (line 128-162) > > According to the measured result by > [https://arthas.aliyun.com/doc/trace.html#id1]) below: Quering deployment may > cost more than 50% (203/349 ms) of time. > Can we save the deployment reference in memory instead of querying it each > time to save pod creating time? > !trace.png! -- This message was sent by Atlassian Jira (v8.20.1#820001)
[jira] [Commented] (FLINK-25648) Redundant to query deployment when creating task manager pods
[ https://issues.apache.org/jira/browse/FLINK-25648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17498699#comment-17498699 ] Yuan Huang commented on FLINK-25648: - Thank you for your time. Let me know if there needs any update. > Redundant to query deployment when creating task manager pods > - > > Key: FLINK-25648 > URL: https://issues.apache.org/jira/browse/FLINK-25648 > Project: Flink > Issue Type: Improvement > Components: Deployment / Kubernetes >Affects Versions: 1.14.2 >Reporter: Yuan Huang >Assignee: Yuan Huang >Priority: Critical > Labels: pull-request-available > Attachments: trace.png > > > When creating Task Manager (TM) Pod, the fabric client needs the deployment > to set owner reference for the TM pod. However, repeating to querying the > deployment each time is unnecessary which will waste a lot of time. > [https://github.com/apache/flink/blob/36ff71f5ff63a140acc634dd1d98b2bb47a76ba5/flink-kubernetes/src/main/java/org/apache/flink/kubernetes/kubeclient/Fabric8FlinkKubeClient.java] > (line 128-162) > > According to the measured result by > [https://arthas.aliyun.com/doc/trace.html#id1]) below: Quering deployment may > cost more than 50% (203/349 ms) of time. > Can we save the deployment reference in memory instead of querying it each > time to save pod creating time? > !trace.png! -- This message was sent by Atlassian Jira (v8.20.1#820001)
[jira] [Commented] (FLINK-25648) Redundant to query deployment when creating task manager pods
[ https://issues.apache.org/jira/browse/FLINK-25648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17495830#comment-17495830 ] Yang Wang commented on FLINK-25648: --- Thanks for working on this ticket. I will get to the PR in this week. > Redundant to query deployment when creating task manager pods > - > > Key: FLINK-25648 > URL: https://issues.apache.org/jira/browse/FLINK-25648 > Project: Flink > Issue Type: Improvement > Components: Deployment / Kubernetes >Affects Versions: 1.14.2 >Reporter: Yuan Huang >Assignee: Yuan Huang >Priority: Critical > Labels: pull-request-available > Attachments: trace.png > > > When creating Task Manager (TM) Pod, the fabric client needs the deployment > to set owner reference for the TM pod. However, repeating to querying the > deployment each time is unnecessary which will waste a lot of time. > [https://github.com/apache/flink/blob/36ff71f5ff63a140acc634dd1d98b2bb47a76ba5/flink-kubernetes/src/main/java/org/apache/flink/kubernetes/kubeclient/Fabric8FlinkKubeClient.java] > (line 128-162) > > According to the measured result by > [https://arthas.aliyun.com/doc/trace.html#id1]) below: Quering deployment may > cost more than 50% (203/349 ms) of time. > Can we save the deployment reference in memory instead of querying it each > time to save pod creating time? > !trace.png! -- This message was sent by Atlassian Jira (v8.20.1#820001)
[jira] [Commented] (FLINK-25648) Redundant to query deployment when creating task manager pods
[ https://issues.apache.org/jira/browse/FLINK-25648?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17495597#comment-17495597 ] Yuan Huang commented on FLINK-25648: - Hi, [~wangyang0918] , the pull request is ready. You can review it when you have time. > Redundant to query deployment when creating task manager pods > - > > Key: FLINK-25648 > URL: https://issues.apache.org/jira/browse/FLINK-25648 > Project: Flink > Issue Type: Improvement > Components: Deployment / Kubernetes >Affects Versions: 1.14.2 >Reporter: Yuan Huang >Assignee: Yuan Huang >Priority: Critical > Labels: pull-request-available > Attachments: trace.png > > > When creating Task Manager (TM) Pod, the fabric client needs the deployment > to set owner reference for the TM pod. However, repeating to querying the > deployment each time is unnecessary which will waste a lot of time. > [https://github.com/apache/flink/blob/36ff71f5ff63a140acc634dd1d98b2bb47a76ba5/flink-kubernetes/src/main/java/org/apache/flink/kubernetes/kubeclient/Fabric8FlinkKubeClient.java] > (line 128-162) > > According to the measured result by > [https://arthas.aliyun.com/doc/trace.html#id1]) below: Quering deployment may > cost more than 50% (203/349 ms) of time. > Can we save the deployment reference in memory instead of querying it each > time to save pod creating time? > !trace.png! -- This message was sent by Atlassian Jira (v8.20.1#820001)