You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@zookeeper.apache.org by "Jingguo Yao (Jira)" <ji...@apache.org> on 2020/03/29 02:26:00 UTC

[jira] [Created] (ZOOKEEPER-3773) Version argument should be checked at the client side

Jingguo Yao created ZOOKEEPER-3773:
--------------------------------------

             Summary: Version argument should be checked at the client side
                 Key: ZOOKEEPER-3773
                 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3773
             Project: ZooKeeper
          Issue Type: Improvement
          Components: java client
    Affects Versions: 3.5.5
            Reporter: Jingguo Yao


Both setData and delete APIs has an argument version. The Javadoc for both APIs says:
{quote}The call will succeed if such a node exists, and the given version matches the
 node's version (if the given version is -1, it matches any node's versions
{quote}
The Javadoc does not specify the behavior if version is less that -1. Here is the behavior I have observed:
 # If the znode does not, KeeperException.NoNodeException is thrown.
 # Otherwise, KeeperException.BadVersionException is thrown.

And the above behavior also holds for Op.Check, Op.Delete and Op.SetData with a version less than -1.

A check of the source code shows that both exception are resulted by checks on the server side. I suggest to do a check on the client. If version is less than -1, KeeperException.BadVersionException is thrown..



--
This message was sent by Atlassian Jira
(v8.3.4#803005)