[ https://forge.continuent.org/jira/browse/SEQUOIA-984?page=comments#action_14392 ]
Emmanuel Cecchet commented on SEQUOIA-984: ------------------------------------------ Disabling a backend for writes does not make much sense in a RAIDb-1 environment since read would be inconsistent which is against the design of RAIDb-1. A full disable should be used in RAIDb-1. Actually I think that the write-disable option should be dropped altogether. This option is only meant for master/slave environments where stale reads are allowed (and I guess this should work with a ParallelDB load balancer). What do you think? > write-disabled backend still accept writes > ------------------------------------------ > > Key: SEQUOIA-984 > URL: https://forge.continuent.org/jira/browse/SEQUOIA-984 > Project: Sequoia > Type: Bug > Components: Core > Versions: sequoia 2.10.10, Sequoia 2.10.9 > Reporter: Stefan Lischke > Assignee: Emmanuel Cecchet > Priority: Critical > > > disabling write for a backend via DatabaseBackend.disableWrite() does not > work. > The backend still accepts writes. > I identified the very strange code in DatabaseBackend: > private boolean canAcceptTasks(boolean isPersistentConnection, > long persistentConnectionId) > { > // return isReadEnabled() || isWriteEnabled() || isReplaying(); > boolean acceptTask = state == BackendState.READ_ENABLED_WRITE_DISABLED > || state == BackendState.READ_ENABLED_WRITE_ENABLED > || state == BackendState.READ_DISABLED_WRITE_ENABLED > || state == BackendState.REPLAYING; > ... > Can someone please tell the semantic behind this method and the or-statement? -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://forge.continuent.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira _______________________________________________ Sequoia mailing list [email protected] https://forge.continuent.org/mailman/listinfo/sequoia
