kafka-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gregoire Seux (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (KAFKA-436) Multifetch response size overflow
Date Fri, 03 Aug 2012 07:22:02 GMT

     [ https://issues.apache.org/jira/browse/KAFKA-436?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Gregoire Seux updated KAFKA-436:
--------------------------------

    Description: 
Fetchsize parameter is per partition in fetch requests. 
This makes the size of response unpredicatable and thus might overflow.

(original email : http://mail-archives.apache.org/mod_mbox/incubator-kafka-users/201208.mbox/%3C416A89FBBC95114F8B2D2E91781541E9638B201E%40SRVEX02.criteois.lan%3E
  )

Three propositions : 
- fetch size is only a hint, broker will ensure that it won't overflow (maybe not sending
too much data)
- broker respect fetchsize as a global limit and not by partition
- broker may not send data for all topics to avoid the overflow.

Changing size to int64 could also decrease probablity of such event.



  was:
Fetchsize parameter is per partition in fetch requests. 
This makes the size of response unpredicatable and thus might overflow.

(original email : http://mail-archives.apache.org/mod_mbox/incubator-kafka-users/201208.mbox/%3C416A89FBBC95114F8B2D2E91781541E9638B201E%40SRVEX02.criteois.lan%3E
  )

Three propositions : 
- fetch size is only a hint, broker will ensure that it won't overflow (maybe not sending
too much data)
- broker respect fetchsize as a global limit and not by partition
- broker may not send data for all topics to avoid the overflow.


    
> Multifetch response size overflow
> ---------------------------------
>
>                 Key: KAFKA-436
>                 URL: https://issues.apache.org/jira/browse/KAFKA-436
>             Project: Kafka
>          Issue Type: Bug
>          Components: core
>    Affects Versions: 0.8, 0.7.1
>            Reporter: Gregoire Seux
>
> Fetchsize parameter is per partition in fetch requests. 
> This makes the size of response unpredicatable and thus might overflow.
> (original email : http://mail-archives.apache.org/mod_mbox/incubator-kafka-users/201208.mbox/%3C416A89FBBC95114F8B2D2E91781541E9638B201E%40SRVEX02.criteois.lan%3E
  )
> Three propositions : 
> - fetch size is only a hint, broker will ensure that it won't overflow (maybe not sending
too much data)
> - broker respect fetchsize as a global limit and not by partition
> - broker may not send data for all topics to avoid the overflow.
> Changing size to int64 could also decrease probablity of such event.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message