[ 
https://issues.apache.org/jira/browse/ZOOKEEPER-1777?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13804355#comment-13804355
 ] 

Flavio Junqueira commented on ZOOKEEPER-1777:
---------------------------------------------

Most of what I'll say I have already said, but given that [~phunt] made a 
comment about including it in 3.4, I'll say it again.

I don't think the situation you're describing here is a bug. I understand that 
for some reason this is important to you, so it makes sense to consider 
changes. However, the changes proposed here touch upon pretty core concepts of 
ZooKeeper, and consequently I'm not comfortable accepting any of these changes 
until I understand precisely their impact and if they are even needed. I have 
already mentioned that if there is any problem, it should be with the 
truncation implementation, but now I'm seeing changes to the format of zxids 
and that gives me headaches. 

I would really appreciate if we could keep it for 3.5.0 so that we can think 
carefully about the implications of any change that is proposed here. I haven't 
dedicated the time this issue deserves because I've been focusing on the 3.4.6 
issues.


> Missing ephemeral nodes in one of the members of the ensemble
> -------------------------------------------------------------
>
>                 Key: ZOOKEEPER-1777
>                 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-1777
>             Project: ZooKeeper
>          Issue Type: Bug
>          Components: quorum
>    Affects Versions: 3.4.5
>         Environment: Linux, Java 1.7
>            Reporter: Germán Blanco
>            Assignee: Germán Blanco
>            Priority: Critical
>             Fix For: 3.5.0
>
>         Attachments: logs_trunk.tar.gz, snaps.tar, ZOOKEEPER-1777-3.4.patch, 
> ZOOKEEPER-1777.patch, ZOOKEEPER-1777.patch, ZOOKEEPER-1777.tar.gz
>
>
> In a 3-servers ensemble, one of the followers doesn't see part of the 
> ephemeral nodes that are present in the leader and the other follower. 
> The 8 missing nodes in "the follower that is not ok" were created in the end 
> of epoch 1, the ensemble is running in epoch 2.



--
This message was sent by Atlassian JIRA
(v6.1#6144)

Reply via email to