ode-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sean Ahn (JIRA)" <j...@apache.org>
Subject [jira] Updated: (ODE-680) Do not send back the response right away when a transaction fails
Date Wed, 14 Oct 2009 21:20:31 GMT

     [ https://issues.apache.org/jira/browse/ODE-680?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Sean Ahn updated ODE-680:
-------------------------

    Description: 
ODE engine currently reschedules a job if the job fails. If a database transaction fails due
to a deadlock, in many cases, the job succeeds on the second try. However, we're sending back
the response right away after the first try.

1. Do not send back the response if the transaction is being rolled back.
2. Send back a failure response on the mex time out. If the mex time out value is larger than
the client time out, the client will get a time out on the response. But, during development,
people can get a better response by controlling the mex time out value.

  was:
ODE engine currently reschedules a job if the job fails in most cases. If a database transaction
fails due to a deadlock, in many cases the job success on the second try. However, we're sending
back the response right away after the first try.

1. Do not send back the response if the transaction is being rolled back.
2. Send back a failure response on the mex time out. If the mex time out value is larger than
the client time out, the client will get a time out on the response. But, during development,
people can get a better response by controlling the mex time out value.


> Do not send back the response right away when a transaction fails
> -----------------------------------------------------------------
>
>                 Key: ODE-680
>                 URL: https://issues.apache.org/jira/browse/ODE-680
>             Project: ODE
>          Issue Type: Improvement
>            Reporter: Sean Ahn
>            Assignee: Sean Ahn
>            Priority: Minor
>
> ODE engine currently reschedules a job if the job fails. If a database transaction fails
due to a deadlock, in many cases, the job succeeds on the second try. However, we're sending
back the response right away after the first try.
> 1. Do not send back the response if the transaction is being rolled back.
> 2. Send back a failure response on the mex time out. If the mex time out value is larger
than the client time out, the client will get a time out on the response. But, during development,
people can get a better response by controlling the mex time out value.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message