[ https://issues.apache.org/jira/browse/BOOKKEEPER-505?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Sijie Guo updated BOOKKEEPER-505: --------------------------------- Attachment: multiple_threads_delivery_manager.patch quite straightforward a patch just move some common code into a 'DeliveryWorker' structure, and change HashMap to ConcucrrentHashMap, and partition the requests by topic. > FIFODeliveryManager runs at single thread. > ------------------------------------------ > > Key: BOOKKEEPER-505 > URL: https://issues.apache.org/jira/browse/BOOKKEEPER-505 > Project: Bookkeeper > Issue Type: Bug > Components: hedwig-server > Reporter: Sijie Guo > Fix For: 4.3.0 > > Attachments: multiple_threads_delivery_manager.patch > > > A lots of things needs to run in FIFODeliveryManager, > #startServingSubscription, #stopServingSubscriber, resuming throttled > subscriber and delivering first message for each started subscriber. It would > increment latency for delivering message or sending response for > stopServingSubscriber (response for closeSubscription, or subscription event > when force attach or topic release). > It is similar as what readahead cache happened before. > -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira