[ https://issues.apache.org/jira/browse/HDFS-11091?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15685990#comment-15685990 ]
Yuanbo Liu commented on HDFS-11091: ----------------------------------- [~zhz] Sorry to interrupt. I've read your discussion in HDFS-9799. {{IOException}} in {{getTrashRoot}} and {{getTrashRoots}} is removed because of compatibility consideration. The idea in HDFS-9799 is to swallow exception and fall back to default trash root with warning log, users can aware the trash root is wrong by using {{rm}} afterwards. I tag you here because after discussing in HDFS-10756, we still think it's a surprising behavior to fall back here. I'd propose to throw a run time exception here. If you have any thought about this JIRA, please let me know, thanks in advance. > Implement a getTrashRoot that does not fall-back > ------------------------------------------------ > > Key: HDFS-11091 > URL: https://issues.apache.org/jira/browse/HDFS-11091 > Project: Hadoop HDFS > Issue Type: Improvement > Affects Versions: 2.8.0 > Reporter: Xiao Chen > Assignee: Yuanbo Liu > > From HDFS-10756's > [discussion|https://issues.apache.org/jira/browse/HDFS-10756?focusedCommentId=15623755&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15623755]: > {{getTrashRoot}} is supposed to return the trash dir considering encryption > zone. But if there's an error encountered (e.g. access control exception), it > falls back to the default trash dir. > Although there is a warning message about this, it is still a somewhat > surprising behavior. The fall back was added by HDFS-9799 for compatibility > reasons. This jira is to propose we add a getTrashRoot that throws, which > will actually be more user-friendly. -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org