[
https://issues.apache.org/jira/browse/ZOOKEEPER-920?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12929974#action_12929974
]
Michi Mutsuzaki commented on ZOOKEEPER-920:
-------------------------------------------
Hi Chang,
Thank you for the patch. Here are my comments:
- Could you document the return value of app_ping_fn? Something like "\return 0
on successful ping. non-zero on error" should be enough.
- Do you have a unit test for this?
- It might be better if send_ping returns something other than
ZOPERATIONTIMEOUT in case of client ping failure. That way it's easier to tell
whether it's client ping that failed.
Thanks!
Michi
> L7 (application layer) ping support
> -----------------------------------
>
> Key: ZOOKEEPER-920
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-920
> Project: Zookeeper
> Issue Type: New Feature
> Components: c client
> Affects Versions: 3.3.1
> Reporter: Chang Song
> Assignee: Chang Song
> Priority: Minor
> Fix For: 3.4.0
>
> Attachments: ZOOKEEPER-920.patch
>
>
> Zookeeper is used in applications where fault tolerance is important. Its
> client i/o thread send/recv heartbeats to/fro Zookeeper ensemble to stay
> connected. However healthy heartbeat does not always means that the
> application that uses Zookeeper client is in good health, it only means that
> ZK client thread is in good health.
> This I needed something that can tagged onto Zookeeper ping that represents
> L7 (application) health as well.
> I have modified C client source to support this in minimal way. I am new to
> Zookeeper, so please code review this code. I am actually using this code in
> our in-house solution.
> https://github.com/tru64ufs/zookeeper/commit/2196d6d5114a2fd2c0a3bc9a55f4494d47d2aece
> Thank you very much.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
