hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Anoop Sam John (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-3787) Increment is non-idempotent but client retries RPC
Date Mon, 04 Feb 2013 09:22:25 GMT

    [ https://issues.apache.org/jira/browse/HBASE-3787?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13570083#comment-13570083
] 

Anoop Sam John commented on HBASE-3787:
---------------------------------------

Some questions on how it will behave
1. When we get a request from client(retry) and old request is in progress, if we throw exception
how the user will know whether the opeartion was success or not?
2. When we get a request from client(retry) and old request is already completed and operation
was successful, I think we need to make that request a successful one. Ultimately the operation
was successful and the client has no knowledge abt that.
3. When we get a request from client(retry) and old request is already completed and operation
failed, we can go ahead with this retry request? Failure would have been because of some thing
like WAL write failure etc which can be retried any way. So this way there is no need to keep
the nonce for grace period when the operation failed.
When it was success we need to keep for a grace period and then expire.
                
> Increment is non-idempotent but client retries RPC
> --------------------------------------------------
>
>                 Key: HBASE-3787
>                 URL: https://issues.apache.org/jira/browse/HBASE-3787
>             Project: HBase
>          Issue Type: Bug
>          Components: Client
>    Affects Versions: 0.96.0, 0.94.4
>            Reporter: dhruba borthakur
>            Priority: Critical
>             Fix For: 0.96.0
>
>
> The HTable.increment() operation is non-idempotent. The client retries the increment
RPC a few times (as specified by configuration) before throwing an error to the application.
This makes it possible that the same increment call be applied twice at the server.
> For increment operations, is it better to use HConnectionManager.getRegionServerWithoutRetries()?
Another  option would be to enhance the IPC module to make the RPC server correctly identify
if the RPC is a retry attempt and handle accordingly.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message