You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@zookeeper.apache.org by "wanglei (Jira)" <ji...@apache.org> on 2019/08/30 05:09:00 UTC
[jira] [Updated] (ZOOKEEPER-3526) When a zk Node just become the leader,synchronizing with Follower in the first time, the value of maxCommittedLog may be smaller than the value of minCommittedLog，is that a problem?
[ https://issues.apache.org/jira/browse/ZOOKEEPER-3526?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
wanglei updated ZOOKEEPER-3526:
-------------------------------
Description:
1.Node1:1566815238 (myid)、Node2:1566815239 (myid)、Node3:1566815240 (myid)
2.After a election, Node3 become the new leader, begin to sync with followers
!image-2019-08-30-12-57-33-883.png!
maxCommittedLog=0x3 minCommittedLog=0x9000002d9 peerLastZxid=0x9000004ca
*why maxCommittedLog < minCommittedLog?*
2.Node 2(follower) get a Trunc message form leader.The leader zxid of the Trunc message is 0x3. So Node3 truncat the transaction log(the zxid which is bigger than 0x3 will be deleted). At last, the data in Node2 is inconsistent.
!image-2019-08-30-13-07-31-467.png!
was:
1.Node1:1566815238 (myid)、Node2:1566815239 (myid)、Node3:1566815240 (myid)
2.After a election, Node3 become new leader，
> When a zk Node just become the leader,synchronizing with Follower in the first time, the value of maxCommittedLog may be smaller than the value of minCommittedLog，is that a problem?
> --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: ZOOKEEPER-3526
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3526
> Project: ZooKeeper
> Issue Type: Bug
> Components: server
> Affects Versions: 3.4.14
> Environment: 1.version:3.4.14
> 2.num of zk nodes: 3
> Reporter: wanglei
> Priority: Major
>
> 1.Node1:1566815238 (myid)、Node2:1566815239 (myid)、Node3:1566815240 (myid)
> 2.After a election, Node3 become the new leader, begin to sync with followers
> !image-2019-08-30-12-57-33-883.png!
> maxCommittedLog=0x3 minCommittedLog=0x9000002d9 peerLastZxid=0x9000004ca
> *why maxCommittedLog < minCommittedLog?*
> 2.Node 2(follower) get a Trunc message form leader.The leader zxid of the Trunc message is 0x3. So Node3 truncat the transaction log(the zxid which is bigger than 0x3 will be deleted). At last, the data in Node2 is inconsistent.
> !image-2019-08-30-13-07-31-467.png!
>
--
This message was sent by Atlassian Jira
(v8.3.2#803003)