You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "ASF GitHub Bot (Jira)" <ji...@apache.org> on 2023/01/27 11:45:00 UTC
[jira] [Updated] (HIVE-26600) Handle failover during optimized bootstrap
[ https://issues.apache.org/jira/browse/HIVE-26600?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
ASF GitHub Bot updated HIVE-26600:
----------------------------------
Labels: pull-request-available (was: )
> Handle failover during optimized bootstrap
> ------------------------------------------
>
> Key: HIVE-26600
> URL: https://issues.apache.org/jira/browse/HIVE-26600
> Project: Hive
> Issue Type: Bug
> Reporter: Teddy Choi
> Assignee: Rakshith C
> Priority: Blocker
> Labels: pull-request-available
> Time Spent: 10m
> Remaining Estimate: 0h
>
> when the reverse policy is enabled from DR to PROD, there is a situation wherein the user may initiate a failover from DR to PROD before the optimized bootstrap is ever run.
> Current observations:
> * Repl Dump will place a failover ready marker but failover metadata won't be generated.
> * Repl Load will throw an error since failover will be set to true but failovermetadata is missing.
> Replication fails and we reach an undefined state.
> Fix :
> * create failover ready marker only during second cycle of optimized bootstrap if possible.
> * since some tables may need to be bootstrapped, it may take upto 3 cycles before failover from DR to PROD is complete.
> * if no tables are modified, second dump from DR to PROD will be marked as failover ready.
> Result:
> * users can initiate a failover immediately after enabling reverse policy without any hassles.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)