You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@zookeeper.apache.org by "fanyang (Jira)" <ji...@apache.org> on 2024/01/29 12:02:00 UTC

[jira] [Created] (ZOOKEEPER-4796) Requests submitted first may carry a larger xid resulting in ZRUNTIMEINCONSISTENCY

fanyang created ZOOKEEPER-4796:
----------------------------------

             Summary: Requests submitted first may carry a larger xid resulting in ZRUNTIMEINCONSISTENCY
                 Key: ZOOKEEPER-4796
                 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-4796
             Project: ZooKeeper
          Issue Type: Bug
          Components: c client
            Reporter: fanyang


When multiple threads attempt to submit requests, it's possible for a request from a thread that acquired its xid earlier to be inserted after a request from a thread that acquired its xid later in the submission queue, which causes a ZRUNTIMEINCONSISTENCY error.

To fix it, acquires the lock before get_xid() and releases it after request submission.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)