zookeeper-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "TisonKun (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (ZOOKEEPER-3289) Throw KeeperException with path in DataTree operations
Date Tue, 26 Feb 2019 07:40:00 GMT

    [ https://issues.apache.org/jira/browse/ZOOKEEPER-3289?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16777681#comment-16777681

TisonKun commented on ZOOKEEPER-3289:

After digging our codes, I think the root of this issue isn't code snippet above. Since we
update {{DataTree}} after the request succeeded.

Assume we execute the follow statements

ZooKeeper zk = ...;
  Op.check(path1, -1),
  Op.delete(path2, -1)));

If path1 or path2 didn't exist, when we executed {{PrepRequestProccessor#pRequest}} in {{PrepRequestProccessor#L804}},
it processed {{KeeperException.NoNodeException}} which contained path info. However, we generated
{{ErrorTxn}} which only contains {{err}} field represented error code. Maybe a reasonable
resolution is extend {{ErrorTxn}} to contain path info or a general {{data}} byte array.

> Throw KeeperException with path in DataTree operations
> ------------------------------------------------------
>                 Key: ZOOKEEPER-3289
>                 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3289
>             Project: ZooKeeper
>          Issue Type: Improvement
>          Components: server
>    Affects Versions: 3.5.4, 3.4.13
>            Reporter: TisonKun
>            Priority: Major
>             Fix For: 3.5.5, 3.4.14
> Currently, if ZooKeeper delete a znode that does not exist. It throws a {{KeeperException.NoNodeException}}
without path message. It causes difficulty when user debug with ZooKeeper. For example,
> Assume we try to do a transaction(with Curator encapsulation)
> {code:java}
> client.inTransaction()
>    .check().forPath(path1).and()
>    .delete().forPath(path2).and()
>    .commit()
> {code}
> if the statement throw an exception {{KeeperException.NoNodeException}} without path
information, we can hardly know that it failed at {{check}} or {{delete}}.
> Thus I propose throws KeeperException with path in DataTree operations. We can achieve
this without burden.

This message was sent by Atlassian JIRA

View raw message