hycdong commented on issue #838: URL: https://github.com/apache/incubator-pegasus/issues/838#issuecomment-971268655
In pegasus, the leader vote is handled by meta server, meaning which replica is primary and which is secondary is determined by meta server, not the replica group. In step2, primary will try to downgrade itself, and it will send a downgrade request to meta server, after that its status is not primary, but a status called inactive. Then, the meta server will choose a suitable secondary and a new primary. And then, meta server will send proposal to vote primary and add secondary to let replica group healthy(1 primary and 2 secondaries). The word in the picture simplifies the process, the request sent to meta is actually NOT the new PartitionConfiguration without the failed secondary, but the request downgrade primary itself. And if the primary failed while sending request, it doesn't matter, and meta server will know this replica group unhealthy in other ways. Hoping my answer helps you. -- 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]
