apurtell commented on PR #4788: URL: https://github.com/apache/hbase/pull/4788#issuecomment-1269148957
> From my perspective, the TaskGroup can be seen as a big MonitoredTask, which contains a list of sub-monitoredTasks. The monitored tasks in the group are still be managed by the TaskMonitor, but whether to clear/expire the monitored tasks in a task group is optional. Since the monitored task already has journals, which mark the phases in a task, we still also need a task group to monitor a big task/process because I think the journals in a task is serial but the tasks in the task group can be parallel(though haven't yet), then we have more flexible ability to monitor the process. That's the idea, yes. Grouping the tasks is not strictly necessary but in my opinion is cleaner for presentation to operators, and also is similar to how Hadoop designed the startup page for the NameNode. We might want to display the tasks in a group in a list view where each task can be collapsed (probably by default) or expanded. See HDFS-4372 and HDFS-4374, upon which our issue was originally inspired. In https://img-blog.csdnimg.cn/929a86ac0f1748f5af70d27ce14f40ee.png "Loading image" is a task group, and "Loading edits" is another task group. Although not strictly required this kind of grouping can help keep the status page layout clean. If for example in our startup status we have a task group "Deploying <tablename>" for each table, the individual items in the group would be the TRSP for each region of <tablename>, and there might be a lot of them, so collapsing the group detail by default would be good. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@hbase.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
