-
Notifications
You must be signed in to change notification settings - Fork 9.2k
HDDS-1490. Support configurable container placement policy through 'o… #903
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
|
💔 -1 overall
This message was automatically generated. |
|
Double checked all failed UT. Two failure reasons,
@xiaoyuyao Do you have a moment to review this latest 04.patch? |
|
@xiaoyuyao, thanks for the information. I didn't releasize that it's because I explicitely added the testResouces to the pom.xml, then the resources under ${basedir}/src/test/resources get ignored. I will upload a new commit shortly. |
…zone.scm.container.placement.classname'
|
💔 -1 overall
This message was automatically generated. |
|
Failed unit tests are all native ACL related. Cannot reproduce them locally. |
xiaoyuyao
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
+1, I will merge it shortly.
… active container failover This PR makes the following changes: * Adds a map called standbyContainerConstraints which stores standby constraints for each container in the job model. The logic for populating the map using the job model is added as Util class. * Adds a check before runStreamProcessor, where the standbyContainerConstraints is checked before launching, if the check succeeds container is launched. If standbyTasks are not enabled in config, existing behaviour is retained. * Adds logic to handle standbyContainerConstraints check failures, which entails releasing the resource, and making an any-host request for the container. * Adds logic in SamzaResourceRequest to order container request such that active-containers take precedence over standby, otherwise based on request-timestamp. * Adds logic in HostAwareContainerAllocator to trap resourceRequests issued by the CPM and the HACA, and a. if it is an any-host for an active container, translate it for a stop a standby. b. if it is for a standby stopped by us, translate it for a resource request for active on standby-host and standby on anyhost. c. in all other cases, proceeds asis. * Adds metrics to capture FailedStandbyAllocations (due to standby constraints), successful standby allocations, number of failovers of active to a valid standby, number of failovers of active to anyhost, in case no standby was found, and number of standby-container-stops completed. Tested on dev setup, a VPC, and a test cluster. Author: Ray Matharu <[email protected]> Author: rmatharu <[email protected]> Reviewers: Jagadish Venkatraman <[email protected]> Closes apache#903 from rmatharu/test-basicStandbyFailover

…zone.scm.container.placement.classname'