[
https://issues.apache.org/jira/browse/ZOOKEEPER-784?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12901028#action_12901028
]
Flavio Junqueira commented on ZOOKEEPER-784:
--------------------------------------------
Nice job, Sergey! It is pretty well done, so I just have a few comments:
# It is worth adding a warn message to ZooKeeperServer to log when an old
client connects to an ro-enabled server. Currently the catch block doesn't do
anything;
# The code of ReadOnlyRequestProcessor is straightforward, but it would be nice
to add a high-level description to the beginning of the class (add javadocs in
general);
# I wonder if waiting to start the r-o server for a single tickTime is not too
aggressive. Would it make sense to set a lower bound of say 2s? That is, we use
the largest between tickTime and 2s. An alternative is to add an optional
parameter to set this value;
# I couldn't really understand how the r-o zookeeper server gets killed once
the replica joins a working quorum (i.e., once it transitions to either
FOLLOWING, LEADING, or OBSERVING). I understand that leader election keeps
running regularly, and the ro server starts if the state of the server doesn't
change within the timeout period, but I couldn't find the code that kills the
r-o server in the case it starts and the replica is not LOOKING. Could you
please point me out to code that does it?
> server-side functionality for read-only mode
> --------------------------------------------
>
> Key: ZOOKEEPER-784
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-784
> Project: Zookeeper
> Issue Type: Sub-task
> Components: server
> Reporter: Sergey Doroshenko
> Assignee: Sergey Doroshenko
> Fix For: 3.4.0
>
> Attachments: ZOOKEEPER-784.patch, ZOOKEEPER-784.patch,
> ZOOKEEPER-784.patch, ZOOKEEPER-784.patch, ZOOKEEPER-784.patch,
> ZOOKEEPER-784.patch, ZOOKEEPER-784.patch, ZOOKEEPER-784.patch,
> ZOOKEEPER-784.patch
>
>
> As per http://wiki.apache.org/hadoop/ZooKeeper/GSoCReadOnlyMode , create
> ReadOnlyZooKeeperServer which comes into play when peer is partitioned.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
