[
https://issues.apache.org/jira/browse/ZOOKEEPER-1777?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Germán Blanco updated ZOOKEEPER-1777:
-------------------------------------
Attachment: ZOOKEEPER-1777.patch
For my case there is a simple solution, since our snapshots are very small we
have already applied a patch that forces snapshot synchronization and avoids
the problem. In any case, severity was changed by Patrick Hunt, you may want to
check with him in case you haven't done so already.
The attached patch proposes a fix in which an incremental hash that should be
unique for each transaction history is associated with each transaction. This
hash is sent to the Leader (only if the leader supports it).
The Leader then sends an snapshot if the hash doesn't match its history for the
same transaction.
At least this was the intention of the change :-).
I had only time to check the patch for 3.4 and at least it passes the
regression test.
Reviews and comments will be very appreciated.
> 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: Blocker
> Fix For: 3.4.6, 3.5.0
>
> Attachments: logs_trunk.tar.gz, snaps.tar, ZOOKEEPER-1777-3.4.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)
