vinayakumarb commented on PR #6792: URL: https://github.com/apache/hadoop/pull/6792#issuecomment-2108716673
> @zhuzilong2013 Thanks for your report and contribution! IMO, they are independent between different BPServiceActor, if exit DN process due to one BPServiceActor issue, it will increase number of Dead DataNode from the whole cluster view, where I don't think it is proper in Federation Arch. Another side, maybe we could add some BPServiceActor count metric to monitor if BPServiceActor works fine? Thanks again. +1 One BPServiceActor reports to one namenode. In case of HA, if one of the namenode not able to connect due to some reason, DN can continue to report to available NameNode. Morever, if all BPServiceActor of a BPOfferService (i.e. connections to all namenodes belonging to same namespace) exited, BPOfferService also shutsdown. When all of such BPOfferServices (in case of federation, there will be multiple) shutdown, Datanode will automatically initiate the shutdown. Refer `DataNode.join()` method. -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
