[
https://issues.apache.org/jira/browse/HBASE-15734?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15264290#comment-15264290
]
Stephen Yuan Jiang commented on HBASE-15734:
--------------------------------------------
[~tedyu], when you do setFailure, the first argument is for the source of the
exception. I see you put the phase of the procedure in the source; usually
based on exception, it is clear which phase failed. If you think it is not
clear where the failure is, maybe you can add the procedure name before the
phase (eg. something like (just an example) "master-full-backup: pre-snapshot
phase" - no need mention that is from a failure). Just my suggestion, it is up
to you to take it or not. Overall, the patch should fix the problem in this
JIRA.
> Propagate exception in table backup procedures
> ----------------------------------------------
>
> Key: HBASE-15734
> URL: https://issues.apache.org/jira/browse/HBASE-15734
> Project: HBase
> Issue Type: Bug
> Reporter: Ted Yu
> Assignee: Ted Yu
> Labels: backup
> Attachments: 15734.v1.txt
>
>
> Currently we have the following in XXTableBackupProcedure:
> {code}
> } catch (Exception e) {
> // fail the overall backup and return
> failBackup(env, backupInfo, backupManager, e, "Unexpected
> BackupException : ",
> BackupType.FULL, conf);
> return Flow.NO_MORE_STATE;
> {code}
> However, failBackup() doesn't propagate the exception to procedure V2.
> This issue is to add setFailure() calls for the propagation.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
