[ 
https://issues.apache.org/jira/browse/ACE-381?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13747445#comment-13747445
 ] 

Ronald Spierenburg commented on ACE-381:
----------------------------------------

The current store implementation can create multiple (log-store) files, 
rotating to a new one when an exception occurs while working with the current 
one. There is no support for maximum filesizes etc, the file will keep on 
growing. 

Because the feedbackchannels will be initialized from a configuration it makes 
sense to add the maximum size to this configuration. The files backing this 
feedbackchannel will be split according to the maximum size to make sure that 
each file will be <10% of the maximum.

When a new feedbackchannel logevent is written we will check if the file should 
be rotated. If the file should rotate then check if all the files on disk are 
greater than the maximum size for this channel. When the maximum is reached 
remove old files until a certain threshold is reached (for example 90% of the 
maximum size). 
This cleanup operation will be fast enough to keep the implementation 
single-threaded as is it now.

The storeId for a log must be reused for a long as possible (because of the way 
the server processes these logs to search for successful deployments). 

The proposed solution : Add the max filesize (and perhaps other usefull 
properties) to the feedbackchannel configuration, create multiple log-files 
<10% of maxSize while reusing the storeId, remove old files when maximum is 
reached.
                
> Truncate target logs in a configurable way.
> -------------------------------------------
>
>                 Key: ACE-381
>                 URL: https://issues.apache.org/jira/browse/ACE-381
>             Project: ACE
>          Issue Type: Improvement
>          Components: Log, Management Agent
>            Reporter: Marcel Offermans
>            Assignee: Angelo van der Sijpt
>
> Target logs should be truncated to a fixed maximum size to prevent disk 
> storage from overflowing.

--
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

Reply via email to